Logging query language

This document describes, at a high level, the Logging query language that you use to query and filter Cloud Logging data.

For in-depth information about the Logging query language design, see the Google API formal specifications for filtering.

For examples of common queries you might want to use, see Sample queries using the Logs Explorer.

Overview

You can use the Logging query language in the Logs Explorer in the Google Cloud console, the Logging API, or the command-line interface. You can use the Logging query language to query data and to write filters to create sinks and log-based metrics.

A query is a Boolean expression that specifies a subset of all the log entries in your selected Google Cloud resource, such as a Google Cloud project or folder.

You can build queries based on the LogEntry indexed field using the logical operators AND and OR. Using the resource.type field in the following examples, the Logging query language grammar looks like this:

  • Simple restriction: resource.type = "gae_app"

  • Conjunctive restriction: resource.type = "gae_app" AND severity = "ERROR"

  • Disjunctive restriction: resource.type = "gae_app" OR resource.type = "gce_instance"

    • Alternatively: resource.type = ("gae_app" OR "gce_instance")
  • Complex conjunctive/disjunctive expression: resource.type = "gae_app" AND (severity = "ERROR" OR "error")

Following is a simple example of a query:

resource.type = "gce_instance" AND
severity >= "ERROR" AND
NOT textPayload:robot

This query matches log entries from Compute Engine that have severity values of at least ERROR and whose textPayload field doesn't contain the string robot anywhere inside it. String comparisons aren't case sensitive. The names resource, severity, and textPayload are defined in the LogEntry type.

Syntax notation

The following sections provide an overview of the Logging query language syntax, and discuss in detail how queries are structured and how matching is performed. Some of the examples use comments to provide explanatory text.

Note the following:

  • The length of a query can't exceed 20,000 characters.

  • The Logging query language is case-insensitive, with the exception of regular expressions.

Syntax summary

The Logging query language syntax can be thought of in terms of queries and comparisons.

A query is a string containing an expression:

expression = ["NOT"] comparison { ("AND" | "OR") ["NOT"] comparison }

A comparison is either a single value or a Boolean expression:

"The cat in the hat"
resource.type = "gae_app"

The first line is an example of a comparison that is a single value. These types of comparisons are global restrictions. Each field of a log entry is compared to the value by implicitly using the has operator. For this example, if any field in a LogEntry, or if its payload, contains the phrase "The cat in the hat", then the comparison is successful.

The second line is an example of a comparison that is a Boolean expression of the form [FIELD_NAME] [OP] [VALUE]. The elements of the comparison are described below:

  • [FIELD_NAME] is a field in a log entry. For example, resource.type.

  • [OP] is a comparison operator. For example, =.

  • [VALUE] is a number, string, function, or parenthesized expression. For example, "gae_app". For JSON null values, use NULL_VALUE.

Boolean operators

The Boolean operators AND and OR are short-circuit operators. The NOT operator has the highest precedence, followed by OR and AND in that order. For example, the following two expressions are equivalent:

"a" OR NOT "b" AND NOT "c" OR "d"
("a" OR (NOT "b")) AND ((NOT "c") OR "d")

You can omit the AND operator between comparisons. You can also replace the NOT operator with the - (minus) operator. For example, the following two queries are the same:

a="b" AND c="d" AND NOT e="f"
a="b" c="d" -e="f"

This documentation always uses AND and NOT.

For all filters except filters used by log views, you can use AND, OR, and NOT operators. Log views only support AND and NOT operations.

To combine AND and OR rules in the same expression, you must nest the rules using parentheses. If you don't use parentheses, your query might not work as intended.

Boolean operators always need to be capitalized. Lowercase and, or, and not are parsed as search terms.

Comparisons

Comparisons have the following form:

[FIELD_NAME] [OP] [VALUE]

The elements of the comparison are described below:

  • [FIELD_NAME]: is the path name of a field in a log entry. Examples of the field name are:

    resource.type
    resource.labels.zone
    resource.labels.project_id
    insertId
    jsonPayload.httpRequest.protocol
    labels."compute.googleapis.com/resource_id"
    

    If a component of a path name has special characters, the path name needs to be double-quoted. For example, compute.googleapis.com/resource_id needs to be double quoted because it contains a forward slash /.

    For details, see field path identifiers in this document.

  • [OP]: is a comparison operator, one of the following:

    =           -- equal
    !=          -- not equal
    > < >= <=   -- numeric ordering
    :           -- "has" matches any substring in the log entry field
    =~          -- regular expression search for a pattern
    !~          -- regular expression search not for a pattern
    

To learn how to search log entries using regular expressions, see Using regular expressions.

  • [VALUE]: is a number, string, function, or parenthesized expression. Strings are used to represent arbitrary text, plus Boolean, enumeration, and byte-string values. The [VALUE] is converted to the field's type prior to the comparison. For JSON null values, use NULL_VALUE.

To filter for a JSON null value, use the following syntax:

jsonPayload.field = NULL_VALUE      -- includes "field" with null value
NOT jsonPayload.field = NULL_VALUE  -- excludes "field" with null value

If [VALUE] is a parenthesized Boolean combination of comparisons, then the field name and the comparison operator are applied to each element. For example:

jsonPayload.cat = ("longhair" OR "shorthair")
jsonPayload.animal : ("nice" AND "pet")

The first comparison checks that the field cat has the value "longhair" or "shorthair". The second checks that the value of the field animal contains both of the words "nice" and "pet", in any order.

Field path identifiers

All log entries are instances of type LogEntry. The identifier that is (or begins) the left-hand side of a comparison must be a field defined in the LogEntry type. For details on the possible identifiers and their values, see the LogEntry type.

Here is the current list of log entry fields. Each field is followed by the next level of names for that field, if applicable:

  • httpRequest: { cacheFillBytes, cacheHit, cacheLookup, cacheValidatedWithOriginServer, latency, protocol, referer, remoteIp, requestMethod, requestSize, requestUrl, responseSize, serverIp, status, userAgent }
  • insertId
  • jsonPayload { variable }
  • labels { variable }
  • logName
  • metadata { systemLabels, userLabels }
  • operation{ id, producer, first, last }
  • protoPayload { @type, variable }
  • receiveTimestamp
  • resource { type, labels }
  • severity
  • sourceLocation: { file, line, function }
  • spanId
  • textPayload
  • timestamp
  • trace

Following are examples of field path identifiers you can use in your comparisons:

  • resource.type: If your first path identifier is resource, then the next identifier must be a field in the MonitoredResource type.

  • httpRequest.latency: If your first path identifier is httpRequest, then the next identifier must be a field in the HttpRequest type.

  • labels.[KEY] If your first path identifier is labels, then the next identifier, [KEY], must be one of the keys from the key-value pairs appearing in the labels field.

  • logName: Since the logName field is a string, you can't follow it by any subfield names.

When you query map or struct fields, you must preserve their keys' letter case and formatting in your expression.

For example, jsonPayload is a struct field, so a field name nested inside jsonPayload like jsonPayload.end_time differs from jsonPayload.endTime. Similarly, for a map field like labels, the label key labels.env_name is different than labels.envName. In contrast, when querying the regular protocol buffer field protoPayload, you don't need to preserve case.

For information on the LogEntry field types, see the google.logging.v2 reference.

Special characters

If a LogEntry field contains special characters, the log field must be quoted. For example:

jsonPayload.":name":apple

jsonPayload."foo.bar":apple

jsonPayload."\"foo\"":apple

For the list of special characters, see the string section in Values and conversions.

For more information on using field path identifiers that reference objects or arrays, see Object and array types in this document.

Monitored resource types

For faster queries, specify a monitored resource type. For a list of resource types, see Monitored resource types.

For example, Compute Engine VMs use the resource type gce_instance and Amazon EC2 instances use aws_ec2_instance. The following example shows how to limit your queries to both type of VMs:

resource.type = ("gce_instance" OR "aws_ec2_instance")

The monitored resource type values in logs are indexed. Using substring matches for them results in slower queries.

Missing fields

If you use a field name in a query, and that field doesn't appear in a log entry, then the field is missing, undefined, or defaulted:

  • If the field is part of the log entry's payload (jsonPayload or protoPayload), or if it is in a label in the labels section of the log entry, then the field is missing. Using a missing field won't display an error, but all comparisons using missing fields fail silently.

    Examples: jsonPayload.nearest_store, protoPayload.name.nickname

  • If the field is defined in the LogEntry type, then the field is defaulted. Comparisons are performed as if the field were present and had its default value.

    Examples: httpRequest.remoteIp, trace, operation.producer

  • Otherwise, the field is undefined, which is an error that is detected before the query is used.

    Examples: thud, operation.thud, textPayload.thud

To test if a missing or defaulted field exists without testing for a particular value in the field, use the :* comparison. For example, the following comparison succeeds if the field operation.id is explicitly present in a log entry:

operation.id:*

Note the behavior of the following queries:

  • When you use the Boolean NOT operator on a missing field, the result is TRUE:

    -- Returns TRUE
    NOT missingField=foo
    
  • When you use the not equal comparison operator != on a missing field, the result is FALSE:

    -- Returns FALSE
    missingField!=foo
    

Object and array types

Each log entry field can hold a scalar, object, or array.

  • A scalar field stores a single value, like 174.4 or -1. A string is also considered a scalar. Fields that can be converted to (or from) a string, such as Duration and Timestamp are also scalar types.

  • An object type stores a collection of named values, like the following JSON value:

    {"age": 24, "height": 67}
    

    You can refer to value inside an object. For example, if jsonPayload.x contained the preceding value, then jsonPayload.x.age would have the value 24.

  • An array field stores a list of values—all of the same type. For example, a field holding measurements might have an array of numbers:

    {8.5, 9, 6}
    

    When comparisons are performed and [FIELD_NAME] is an array field, each member of the array is compared to [VALUE] and the results are joined together using the OR operator. For example, if jsonPayload.shoeSize is an array field that stores {8.5, 9, 6}, the comparison:

    jsonPayload.shoeSize < 7
    

    is equivalent to:

    8.5 < 7 OR 9 < 7 OR 6 < 7
    

    In this example, the overall comparison evaluates to successful.

Values and conversions

The first step in evaluating a comparison is to convert the right-hand side value to the type of the log entry field. Scalar field types are permitted in comparisons, along with two additional types whose values are represented as strings: Duration and Timestamp. For a list of scalar types, see the scalar protocol buffer types list. The following table explains what values can be converted to the log field types:

Field type Permitted query value
bool

"True" or "false" in any letter case. Examples: "True", "true"

bytes

A string containing any sequence of bytes. Example: "\377\377".

Duration

A string containing a signed decimal number followed by one of the units "ns", "us", "ms", "s", "m", or "h". Durations are accurate to nanoseconds. Example: "3.2s".

enum

The name of an enumeration type literal, case-insensitive. Examples: "WARNING", which is a value of type LogSeverity.

double

Any number, with or without a sign and an exponent part, or the special value strings "NaN", "-Infinity", and "Infinity" (either capitalized or not). Examples: "-3.2e-8", "nan".

intNN

Any signed integer that doesn't exceed the size of the type. Example: "-3".

string

Any string that contains UTF-8 encoded or 7-bit ASCII text. Embedded quotation marks must be escaped with a backslash.

String values must be double-quoted to escape the following special characters:

  • Strings starting with + (plus), - (minus), or . (period).

  • Strings with ~ (tilde), = (equals), () (parentheses), : (colon), > (greater than), < (less than), , (comma), . (period), or * (asterisk).

  • Any escape sequence, for example \t.

Timestamp

A string in RFC 3339 or ISO 8601 format. Examples: "2014-10-02T15:01:23.045Z" (RFC 3339), "2014-10-02" (ISO 8601). In query expressions, timestamps in RFC 3339 format can specify a timezone with "Z" or ±hh:mm. Timestamps are represented to nanosecond accuracy.

uintNN

Any unsigned integer that doesn't exceed the size of the type. Example: "1234".

If an attempted conversion fails, then the comparison fails.

When a conversion requires a string, you can also use a number or unquoted text if they don't contain special characters such as spaces and operators. Similarly, when a conversion requires a number, you can use a string whose content is a number.

The types intNN and uintNN represent integer types of various sizes, such as int32 and uint64. When writing a value to be converted to a 64-bit integer type, you write the value as a string, such as "9223372036854775807".

Types of log fields

Here is how the type of a log entry field is determined:

  • Log fields defined in the type LogEntry, and in the component type are protocol buffer fields. Protocol buffer fields have explicit types.

  • Log fields that are part of protoPayload objects are also protocol buffer fields and have explicit types. The name of the protocol buffer type is stored in the field "@type" of protoPayload. For more information, see the JSON mapping.

    When you are filtering on a field that is associated with the Any message type, the value field is automatically traversed. Therefore, don't include it in the query. For more information, see Troubleshooting.

  • Log fields inside of jsonPayload have types that are inferred from the field's value when the log entry is received:

    • Fields whose values are unquoted numbers have type double.
    • Fields whose values are true or false have type bool.
    • Fields whose values are strings have type string.

    Long (64-bit) integers are stored in string fields, because they can't be represented exactly as double values.

  • The Duration and Timestamp types are recognized only in protocol buffer fields. Elsewhere, those values are stored in string fields.

Comments

Comments start with two dashes (--), and any text following the dashes is ignored until the end of the line. Comments can be placed at the beginning of a filter, in between terms, and at the end of a filter.

You might use comments for the following cases:

  • To annotate your complex filters with information about what a clause does:

     -- All of our target users are emitted by Compute Engine instances.
     resource.type = "gce_instance"
     -- Looking for logs from "alex".
     jsonPayload.targetUser = "alex"

  • To quickly enable or disable a clause by adding or removing the comment prefix:

     resource.type = "gce_instance"
     -- jsonPayload.targetUser = "alex"
     jsonPayload.targetUser = "kiran"
     -- jsonPayload.targetUser = "sasha"

Comparison operators

The meaning of the equality (=, !=) and inequality (<, <=, >, >=) operators depends on the underlying type of the left-hand field name.

  • All numeric types: Equality and inequality have their normal meaning for numbers.
  • bool: Equality means the same Boolean value. Inequality is defined by true>false.
  • enum: Equality means the same enumeration value. Inequality uses the underlying numeric values of the enumeration literals.
  • Duration: Equality means the same duration length. Inequality is based on the length of the duration. Example: as durations, "1s">"999ms".
  • Timestamp: Equality means the same instant in time. If a and b are Timestamp values, a < b means a is earlier in time than b.
  • bytes: Operands are compared byte by byte, left-to-right.
  • string: Comparisons ignore letter case. Specifically, both operands are first normalized using NFKC_CF Unicode normalization and then use lexicographic comparisons. However, regular expression searches aren't normalized. For more information on searching log entries using regular expressions, see Using regular expressions.

The substring operator (:) is applicable to string and bytes, and is handled like equality except that the right-hand operand need only equal some part of the left-hand field. Substring matches on indexed fields don't take advantage of log indexes.

Global restrictions

If the comparison consists of a single value, it is called a global restriction. Logging uses the has (:) operator to determine if any field in a log entry, or if its payload, contains the global restriction. If it does, then the comparison succeeds.

The simplest query written in terms of a global restriction is a single value:

"The Cat in The Hat"

You can combine global restrictions using the AND and OR operators for a more interesting query. For example, if you want to display all log entries that have a field that contains cat and a field that contains either hat or bat, write the query as:

("cat" AND ("hat" OR "bat"))

In this case, there are three global restrictions: cat, hat and bat. These global restrictions are applied separately and the results are combined, just as if the expression had been written without parentheses.

A global restriction is an easy way to query your logs for a particular value. For example, if you are looking in your activity log for entries containing any mention of GCE_OPERATION_DONE, you can use the following query:

logName = "projects/my-project-id/logs/compute.googleapis.com%2Factivity_log" AND
"GCE_OPERATION_DONE"

Although global restrictions are easy, they can be slow; for more information, see Finding log entries quickly in this document.

Functions

You can use built-in functions as global restrictions in queries:

function = identifier ( [ argument { , argument } ] )

where argument is a value, field name, or a parenthesized expression. The functions are described in the following sections.

log_id

The log_id function returns log entries that match the given [LOG_ID] argument from the logName field:

log_id([LOG_ID])

For example, the following query returns all log entries with a cloudaudit.googleapis.com%2Factivity [LOG_ID]:

log_id("cloudaudit.googleapis.com/activity")

source

The source function matches log entries from a particular resource in the organizations, folders, and Google Cloud projects hierarchy.

The source function doesn't match child resources. For example, using source(folders/folder_123) matches logs from the folder_123 resource, and not logs from the Google Cloud project resources within folder_123.

To query for logs at a particular resource level, use the following syntax:

source(RESOURCE_TYPE/RESOURCE_ID)
Resource Example query
Organization source(organizations/ORGANIZATION_ID)
Folder source(folders/FOLDER_ID)
Google Cloud projects source(projects/PROJECT_ID)

sample

The sample function selects a fraction of the total number of log entries:

sample([FIELD], [FRACTION])

[FIELD] is the name of a field in the log entry, such as logName or jsonPayload.a_field. The value of the field determines whether the log entry is in the sample. The field type must be a string or numeric value. Setting [FIELD] to insertId is a good choice, because every log entry has a different value for that field.

[FRACTION] is the fraction of log entries that have values for [FIELD] to include. It is a number greater than 0.0 and no greater than 1.0. For example, if you specify 0.01, then the sample contains roughly one percent of all log entries that have values for [FIELD]. If [FRACTION] is 1, then all the log entries that have values for [FIELD] are chosen.

Example: The following query returns 25 percent of the log entries from log syslog:

logName = "projects/my-project/logs/syslog" AND sample(insertId, 0.25)

Details: A deterministic algorithm, based on hashing, is used to determine if a log entry is included, or excluded, from the sample. The accuracy of the resulting sample is dependent on the distribution of the hashed values. If the hashed values aren't uniformly distributed, then the resulting sample can be skewed. In the worst case, when [FIELD] always contains the same value, the resulting sample contains either the [FRACTION] of all log entries or no log entries.

If [FIELD] does appear in a log entry, then:

  • A hash of the value is computed.
  • The hashed value, which is a number, is divided by the maximum possible hashed value.
  • If the resulting fraction is less than or equal to [FRACTION], the log entry is included in the sample; otherwise it is excluded from the sample.

If [FIELD] doesn't appear in a log entry, then:

  • If [FIELD] is part of the log entry's payload or labels sections, the log entry isn't selected for the sample, even if [FRACTION] is 1.
  • Otherwise, the log entry is treated as if [FIELD] is in the log entry and the value of [FIELD] is the default value. The default value is determined by the LogEntry type. For more information on missing and defaulted fields, see Missing fields in this document.

To exclude log entries with defaulted fields from the sample, use the field-exists operator, :*. The following query produces a 1 percent sample of log entries that have explicitly supplied a value for field:

field:* AND sample(field, 0.01)

ip_in_net

The ip_in_net function determines if an IP address in a log entry is contained in a subnet. You might use this to tell if a request comes from an internal or external source. For example:

ip_in_net([FIELD], [SUBNET])

[FIELD] is a string-valued field in the log entry that contains an IP address or range. The field can be repeating, in which case only one of the repeated fields has to have an address or range contained in the subnet.

[SUBNET] is a string constant for an IP address or range. It is an error if [SUBNET] isn't a legal IP address or range, as described later in this section.

Example: The following query tests an IP address in the payload of log entries from the log my_log:

logName = "projects/my_project/logs/my_log" AND
ip_in_net(jsonPayload.realClientIP, "10.1.2.0/24")

Details: If, in a log entry, [FIELD] is missing, defaulted, or it does not contain a legal IP address or range, then the function returns false. For more information on missing and defaulted fields, see Missing fields in this document.

Examples of the supported IP addresses and ranges follow:

  • IPv4: 10.1.2.3
  • IPv4 subnet: 10.1.2.0/24
  • CIDR IPv6: 1234:5678:90ab:cdef:1234:5678:90ab:cdef
  • CIDR IPv6 subnet: 1:2::/48

SEARCH function

You can use the built-in SEARCH function to find strings in your log data:

SEARCH([query])
SEARCH([field], [query])

Both forms of the SEARCH function contain a query argument, which must be formatted as a string literal. In the first form, the entire log entry is searched. In the second form, you specify the field in the log entry to search.

You must specify the query field. If this field isn't specified, then an error is returned.

When the SEARCH function is processed, the query string is processed by a text analyzer that splits the string into tokens. Cloud Logging always performs case-insensitive comparisons, even for tokens wrapped with backticks. This behavior differs from that of BigQuery, which preserves case in tokens wrapped with backticks. For information about the analyzer rules, see the BigQuery document Text analyzer rules.

When constructing a search, consider the following:

  • Tokens are case-insensitive. The following functions produce the same results:

    SEARCH("world")
    SEARCH("World")
    

    The previous functions match a log entry when a single field contains the token "world". Because SEARCH performs exact matches and not substring matches, the previous functions don't match a field whose value is "worldwide".

  • If you don't specify the field to search, then the SEARCH function matches a log entry when that log entry contains all tokens. However, the order of tokens doesn't matter and the tokens aren't required to be found in the same field of the log entry.

    The following functions produce the same results, and they match a log entry that contains the tokens "hello" and "world":

    SEARCH("hello world")
    SEARCH("World hello")
    
  • If you specify the field to search, then the SEARCH function only searches that field. A match occurs when that field contains all tokens; however, the order of tokens doesn't matter.

    The following functions produce a match only when the textPayload field contains the tokens "hello" and "world":

    SEARCH(textPayload, "hello world")
    
  • To impose a case-insensitive but exact match on a phrase, enclose the phrase in backticks. For example, the following functions match the string "hello world":

    SEARCH("`hello world`")
    SEARCH("`Hello World`")
    SEARCH("`HELLO WORLD`")
    

    Because backticks are used in the following functions, they produce different results:

    SEARCH("`hello world`")
    SEARCH("`world hello`")
    

The Logging query language supports different ways that you can search your log data. When searching for a string, it is more efficient to use the SEARCH function than to perform a global search or a substring search. However, you can't use use the SEARCH function to match non-text fields. For guidance on performing search operations, see Minimize global and substring searches.

Searching by time

In the interface, you can set specific limits on the date and time of log entries to show. For example, if you add the following conditions to your query, the preview displays exactly the log entries in the indicated 30-minute period and you won't be able to scroll outside of that date range:

timestamp >= "2016-11-29T23:00:00Z"
timestamp <= "2016-11-29T23:30:00Z"

When writing a query with a timestamp, you must use dates and times in the format shown above.

You can also search log entries using timestamp shortcuts. For example, you can enter a date with a comparison operator to get all log entries after a certain day:

timestamp > "2016-11-29"

Using regular expressions

You can use regular expressions to build queries and create filters for sinks, metrics, and wherever log filters are used. You can use regular expressions in the Query builder and with Google Cloud CLI.

A regular expression is a sequence of characters that define a search. The Logging query language uses the RE2 syntax. For a complete explanation of the RE2 syntax, see the RE2 wiki on GitHub.

Regular expression queries have the following characteristics:

  • Only fields of the string type can be matched with a regular expression.

  • String normalization isn't performed; for example, kubernetes isn't considered the same as KUBERNETES. For more information, see the Comparison operators section.

  • Queries are case sensitive and not anchored by default.

  • Boolean operators can be used between multiple regular expressions on the right side of the regular expression comparison operator, =~ and !~.

A regular expression query has the following structure:

Match a pattern:

jsonPayload.message =~ "regular expression pattern"

Does not match a pattern:

jsonPayload.message !~ "regular expression pattern"

The =~ and !~ changes the query to a regular expression query, and the pattern you're trying to match must be within double quotation marks. To query for patterns that contain double quotation marks, escape them using a backslash.

Examples querying logs using regular expressions

Query type Example
Standard query sourceLocation.file =~ "foo"
Query with case-insensitive search labels.subnetwork_name =~ "(?i)foo"
Query containing quotation marks jsonPayload.message =~ "field1=\"bar.*\""
Query using a boolean or labels.pod_name =~ "(foo|bar)"
Query using anchors logName =~ "/my%2Flog$"
Query not matching a pattern labels.pod_name !~ "foo"
Query using boolean operator labels.env =~ ("^prod.*server" OR "^staging.*server")
Query that begins with a value logName =~ "^foo"
Query that ends with a value logName =~ "foo$"

Finding log entries quickly

To find log entries more efficiently, do the following:

  • Query using indexed fields.
  • Minimize the number of log entries that must be searched.

Use indexed fields

Logging always indexes the following LogEntry fields:

You can also add custom indexed fields to any log bucket.

The next sections explain how to use indexed fields to minimize the number of log entries to be searched.

Optimize your queries

Make your searches faster by reducing the number of logs, the number of log entries, or the time span of your searches. Even better, you can reduce all three.

Example: Use the right log name

Specify the log containing the log entries you're interested in. Be sure you know the actual log name by inspecting one of your log entries. For example, the preview shows that there is a log in the Compute Engine section named "activity". On closer inspection of the Admin Activity audit log entries, the log is actually named "cloudaudit.googleapis.com/activity".

The following comparison is incorrect. It doesn't match anything because it uses the wrong log name:

logName = "projects/my-project-id/logs/activity"   -- WRONG!

The following comparison is correct. It chooses log entries from the Admin Activity audit log entries. You must URL-encode the log name, as shown:

logName = "projects/my-project-id/logs/cloudaudit.googleapis.com%2Factivity"

Example: Choose the right log entries

If you know that the log entries you want are coming from a particular VM instance, then specify it. Check for the right label names by inspecting one of the log entries that you want to search for. In the following example, instance_id is one of the indexed labels:

resource.type = "gce_instance" AND
resource.labels.instance_id = "6731710280662790612"
logName = "projects/my-project-id/logs/cloudaudit.googleapis.com%2Factivity"

Example: Choose the right time period

Specify a time period to search in. A quick way of determining useful timestamps in RFC 3339 format is to use the GNU/Linux date command:

$ date --rfc-3339=s
2016-06-27 17:39:00-04:00
$ date --rfc-3339=s --date="3 hours ago"
2016-06-27 14:40:00-04:00
$ date --rfc-3339=s --date="5 hours ago"
2016-06-27 12:40:00-04:00

Use the values of these timestamps in the following queries. To create a timestamp acceptable to Logging, replace the space between the date and time with the letter T.

For example, to search within the last three hours:

timestamp >= "2016-06-27T14:40:00-04:00"

As another example, to search between three and five hours ago:

timestamp >= "2016-06-27T12:40:00-04:00" AND
timestamp <= "2016-06-27T14:40:00-04:00"

Minimize global and substring searches

Avoid the temptation to take shortcuts when typing queries.

Example: Don't use global searches

If you're searching for a log entry with "Hello Kitty" in the payload:

  • Don't use a global search. For one reason, they are all substring searches:

       "Hello Kitty"   -- THIS CAUSES A SLOW SEARCH!
       

  • Do limit the search to a single field, even if you must keep the substring search:

       textPayload:"Hello Kitty"
       

  • Do use an equality test if you can:

       textPayload = "Hello Kitty"
       

  • Do reference individual fields in a payload, if your log entries have structured payloads:

       jsonPayload.my_favorite_cat = "Hello Kitty"
       

  • Do use an indexed field to restrict the search:

       logName = "projects/my-project_id/logs/somelog" AND
       jsonPayload.my_favorite_cat = "Hello Kitty"
       

  • Do use the SEARCH function and specify the complete text to match. The SEARCH function performs a case-insensitive match:

      SEARCH("Hello Kitty")
      

    Don't use the SEARCH function and specify partial text. For example, the following function doesn't match "Hello Kitty".

      SEARCH("Hello Kit")
      

Searching examples

The log entries shown are the ones that match a query. If the Jump to time menu contains a value, then the display scrolls to that point in time. Here are some query examples:

resource.type=gae_app

Finds all App Engine log entries. For a list of resource types, see Monitored resource list.

As you type, the preview suggests completions for fields like resource.type.

resource.type=gae_app AND logName:request_log

Finds log entries for App Engine apps from log names containing request_log. Note several things:

  • The = operator is exact equality. The resource type must be exactly "gae_app" except for letter case.
  • The : operator means "has". The logName field must contain request_log, in any letter case. The actual log name is much longer. Using : might result in slower searches.
  • The two comparisons are joined by AND. You can also use OR, but AND is assumed if you leave out the operator.
resource.type = (gce_instance OR aws_ec2_instance) AND severity >= ERROR

Finds log entries with either of two resource types: Compute Engine VM instance or AWS EC2 VM instance. The log entries must have severity of at least ERROR, which is equivalent to selecting ERROR in the query interface's severity menu.

logName = "projects/[PROJECT_ID]/logs/cloudaudit.googleapis.com%2Factivity"

Finds all the Admin Activity audit log entries in the project [PROJECT_ID]. Audit logs all use the same log name in a project, but have different resource types. The log ID, cloudaudit.googleapis.com/activity must be URL-encoded in the log name. Using equality in the comparison speeds up the search. For more information, see Understanding audit logs.

unicorn

Finds log entries containing unicorn in any field, in any letter case. A search term that isn't part of a field comparison is an "all fields" query.

unicorn phoenix

Finds log entries that contain unicorn in some field and phoenix in some field.

textPayload:(unicorn phoenix)

Finds log entries whose textPayload field contains both unicorn and phoenix in any order—the AND is implicit between the two words.

textPayload:"unicorn phoenix"

Finds log entries whose textPayload field contains the string "unicorn phoenix".

NOT textPayload: "unicorn phoenix"

Finds log entries whose textPayload field does not contain the string "unicorn phoenix". This type of query reduces unwanted log entries.

timestamp >= "2016-11-29T23:00:00Z" timestamp <= "2016-11-29T23:30:00Z"

Finds log entries within a 30-minute period.

Troubleshooting

Syntax issues

If you have problems with your queries' expressions, check the following:

  • Your query obeys the syntax rules, with matched parentheses and quotation marks.

  • Your log entry field names are correctly spelled.

  • Boolean operations are in uppercase letters (AND, OR, NOT).

  • Ensure that you're using NULL_VALUE to represent JSON null values.

  • Boolean expressions as global restrictions or as the right-hand side of comparisons should be parenthesized for clarity. For example, the two queries below look the same, but are not:

    insertId = "ABC-1" OR "ABC-2"  -- ERROR!?
    insertId = ("ABC-1" OR "ABC-2")
    
  • Unquoted text must not contain any special characters. When in doubt, add double quotation marks. For example, the first comparison below is illegal because of the embedded substring operator (:). The comparison must be written with quotation marks:

    insertId = abc:def  -- ILLEGAL!
    insertId = "abc:def"
    
  • The Google Cloud CLI requires the query to be in double quotes. To use double quotes for escaping special characters using the gcloud logging command, wrap the entire query with single quotes instead:

    gcloud logging read 'resource.type=gce_instance AND jsonPayload.message="Stopped Unattended Upgrades Shutdown."'
    gcloud logging read 'timestamp>="2020-06-17T21:00:00Z"'
    

  • When you are filtering on a field that is associated with the Any message type, the value field is automatically traversed. Therefore, don't include value in the query.

    For example, the Status field in an AuditLog message has a details field that is of type google.protobuf.Any. To query the details field, omit the value field when specifying the filter:

    • Do

      protoPayload.status.details.conditionNotMet.userVisibleMessage =~ "Specified reservation.*"
      
    • Don't

      protoPayload.status.details.value.conditionNotMet.userVisibleMessage =~ "Specified reservation.*"