Paid feature
Metrics and monitoring¶
By enabling metrics and monitoring feature you will have access to a set of features that help you monitor your subscription. This bundle include features like custom notifications, insights and the metrics API.
How to enable¶
Metrics and monitoring is now available for all subscriptions with clustered architecture. This is how you can activate the new feature:
Login to the Sesam portal
Select the subscription you want to use
Navigate to Subscription on the left menu
Click on Products tab
Click on “Enable Metrics and monitoring”
Pricing¶
Follow the steps above to check how this feature affect your monthly price.
Metrics API¶
This feature will make the /api/metrics
endpoint available. This API endpoint exposes Prometheus compatible
metrics.
After enabling the feature, you will need to set up a prometheus database, then you can use tools, like Grafana, to create your own dashboards.
Note
You need a JWT token with Admin
role to be able to scrape the endpoint.
Example:
curl -s -H "Authorization: bearer $JWT" "$SESAM_API_URL/metrics"
Note that all the counter metrics consists of two metrics suffixed by _counter and _created.
Service metrics¶
Metric |
Type |
Description |
---|---|---|
|
Enum<healthy|unhealthy|unknown> |
Is the Sesam API up? |
|
Enum<healthy|unhealthy|unknown> |
Are there any failed backups? |
|
Gauge |
Number of restarts since last upgrade |
Pipe metrics¶
Metric |
Type |
Description |
---|---|---|
|
Gauge |
Value is |
|
Gauge |
Value is always |
|
Gauge |
Returns bytes used by pipe |
|
Gauge |
Number of source entities in the pipe queue |
|
Gauge |
Number of untracked entities in the pipe queue |
|
Gauge |
Number of entities in the pipe’s replica queues |
|
Gauge |
Number of non-deleted entity ids in the sink dataset |
|
Gauge |
Number of deleted entity ids in the sink dataset |
|
Gauge |
Number of non-deleted entities in the sink dataset log |
|
Gauge |
Number of deleted entities in the sink dataset log |
|
Gauge |
Number of non-deleted entity ids in the dead-letter dataset |
|
Gauge |
Number of deleted entity ids in the dead-letter dataset |
|
Gauge |
Number of non-deleted entities in the dead-letter dataset log |
|
Gauge |
Number of deleted entities in the dead-letter dataset log |
|
Gauge |
Is the circuit-breaker tripped? 1.0 if tripped otherwise 0.0. |
Pump metrics¶
Note that metrics are not exposed for pipes that are disabled or have pump.mode=”manual” or pump.mode=”off”.
Metric |
Type |
Description |
---|---|---|
|
Gauge |
Runtime in seconds. 0.0 if the pipe has not run yet. |
|
Gauge |
Was the last pump run successful? 1.0 if successful and -1.0 if not successful and 0.0 if the pipe has not run yet. |
|
Counter |
The number of pump-started events. Note: no-op runs are not counted currently. |
|
Counter |
The number of pump-completed events. Note: no-op runs are not counted currently. |
|
Counter |
The number of non-interrupted pump-failed events. |
|
Counter |
The number of interrupted pump-failed events. |
|
Counter |
The number of entities changed |
|
Counter |
The number of entities seen |
|
Counter |
The number of entities compacted away |
|
Counter |
The number of entities deletion tracked |
|
Gauge |
The time when the metrics got scraped by the Prometheus client. The unit is the number of seconds since epoch. |
|
Gauge |
The time when the pump last completed or failed. The unit is the number of seconds since epoch. |
|
Gauge |
The time when the pump is next scheduled to run. The unit is the number of seconds since epoch. |
|
Gauge |
The time when the pump was previously scheduled to run. In practice this is the run before the one scheduled at |
Monitoring¶
Monitoring allows you to see pipe insights and set up custom pipe notifications.
Insights¶
After enabling Metrics and monitoring, you can enable insights in the pipe you want to monitor. You will then have access to charts that show how many entities, errors and latencies there have been for the current pipe during the last 30 days.
To enable insights on a specific pipe:
Navigate to the pipe you want to monitor
Find the Insight tab
Enable insights
Important
Monitoring data collection will start only after you enable insights in a specific pipe.
Notifications¶
Add notification rules to pipes and get alerts when those rules are triggered. You can get notification alerts either in the user-interface or by email.
To see how to use custom notifications, please visit the notification documentation.