- What is the system status dashboard?
- What are you measuring?
- How often is it updated?
- How do I read the Status Dashboard?
- Can developers monitor their individual apps using the system dashboard?
- How is the system status measured?
- How do I get a feed of this information?
- How long do notification messages stay on the dashboard?
- How do I report an issue that I don't see on the dashboard?
What is the system status dashboard?
The System Status Dashboard makes it easier for developers to evaluate and monitor the entire App Engine system by enabling them to measure historical uptime, error rates and latency for each of the major App Engine components. We will also use the dashboard to display up-time related announcements from the App Engine team.
What are you measuring?
Google is measuring the performance of each of App Engine's major components; datastore, serving infrastructure, memcache, user authentication, image manipulation and mail infrastructure. For each of these components we measure uptime, error rates and latency. Some of these components like the datastore have detailed measurements per API call. We'd like developers to understand at a glance the availability and latency of every aspect of their applications.
How often is it updated?
The App Engine System Status Dashboard is a continuous monitoring tool. It is updated every few minutes and reflects the current system status. The dashboard tracks average overall uptime, error rate and latency during each update interval. The dashboard will also display uptime-related announcements from the App Engine team.
How do I read the Status Dashboard?
The status dashboard includes the current status of each component probe, as well as the historic status for the past seven days. These statuses are calculated using 10-minute rolling averages of our probe measurements.
The current status of each probe is listed on the right hand side of the dashboard. A status of 'Normal' means that the probe is operating within normal limits. If the error or latency rate is running noticeably slower than average, the dashboard will display a yellow ('Anomaly') status. At when the error or latency rate indicates there is a potential issue, the dashboard will display a red ('Anomaly') status.
The dashboard also contains columns listing the historical system data, which display, by default, the status for the past week. The icons displayed in a cell will indicate what kind of issues, if any, may have occurred during a given day in the past with a specific probe. A green check mark means the probe's measurements were within normal limits all day. An orange question mark indicates that, at some point during the day, the system experienced anomalies which we are still investigating. A yellow exclamation point means that the component in question experienced a service disruption at some point during that day.
When you click on an icon, there will be a graph displaying the probe performance over the course of that day. Using this, you can compare any issues your app may have experienced over a certain time period, with the performance of the relevant App Engine systems during that time period.
Can developers monitor their individual apps using the system dashboard?
No, developers can not use the System Status Dashboard to monitor their applications individually. However, there are many third-party tools that enable developers to monitor their applications.
How is the system status measured?
The system status dashboard has probes to measure calls to each API, as well as a round trip request to a dynamic HTTP handler. The probes measure the latency of these requests, as well as the error rate.
Latency is the measured in milliseconds to complete the API call or request. Error Rate measures the percentage of failed API calls or requests. The description for each probe contains more specific information on each call that is performed by the probe.
The Mail and Memcache API probes also contain throughput statistics. This is the measure of bandwidth of the calls in kbytes per second.
How do I get a feed of this information?
The system status dashboard has an atom feed to which you can subscribe at https://status.cloud.google.com/feed.atom
How long do notification messages stay on the dashboard?
Most notifications stay on the dashboard for 24 hours. Some messages, like scheduled downtime, will remain on the dashboard for the relevant period of time.
How do I report an issue that I don't see on the dashboard?
If you believe you are experiencing a system issue that you do not see in the dashboard, you should report it on the App Engine Google Group. Experimental: You can also report it in the Administration Console via the Report Production Issue button.