Metrics
saturn run pushes metrics to a Prometheus Pushgateway if prometheusPushgatewayUrl is set.
The job label added to all metrics is saturn.
The following metrics exist:
db_size_bytes¶
Exported by: server
Size of the sqlite database file in bytes.
Use this metric to understand if the database file is growing unexpectedly large.
git_commands_duration_seconds_count¶
Exported by: run, worker
Total number of commands executed by git.
Use this metric to understand how much load Saturn puts on a repository host by, for example, cloning repositories or pushing commits.
git_commands_duration_seconds_sum¶
Exported by: run, worker
Total duration it took for git to execute commands.
Use together with git_commands_duration_seconds_count to calculate the average duration
and understand the repository host is slow.
http_client_requests_total¶
Exported by: run, worker
Total number of requests sent via HTTP clients.
This metric includes HTTP requests sent to GitHub or GitLab.
Useful to understand how much load Saturn puts on an API.
run_finish_time_seconds¶
Exported by: run, worker
Last unix time when the run finished.
Use this metric together with run_start_time_seconds
to understand how long the execution of saturn run took.
run_start_time_seconds¶
Exported by: run, worker
Last unix time when the run started.
Use this metric together with run_finish_time_seconds
to understand how long the execution of saturn run took.
run_task_success¶
Exported by: run, worker
Status of the last run of a task. 1 indicates success. 0 indicates failure.
Use this metric to alert that a task has failed.
sb_server_collector_success¶
Exported by: server
Status of the last metric collection.
1 indicates that the last metric collection succeeded.
0 indicates that an error occurred during metric collection.
Use this metric to alert on the availability of metrics.
sb_server_task_run_success¶
Exported by: server
Status of the last run of a task.
1 indicates that the run finished successfully.
0 indicates that the run failed.
Use this metric to alert on failing runs.
Additional label pairs can be added by setting metricLabels
in a task file.
These labels can be used to, for example, route alerts to the actual owners of a task.