[toolsdb] Add db-level and user-level monitoring
Open, Needs TriagePublic
Actions

Assigned To

None

Authored By

	fnegri
	Wed, Jun 3, 5:10 PM

Description

We currently have two Grafana dashboards for ToolsDB:

They include many useful metrics, but they don't include any metrics that are differentiating by ToolsDB users, or ToolsDB database (e.g. number of queries per user or per database). This would be very helpful when debugging spikes in the usage of ToolsDB, to identify which users or tools are responsible for the spike.

Some per-user and per-database metrics can be added to Prometheus by enabling the Performance Schema in MariaDB and in the mysqld_exporter.

Related Objects

Mentioned In: T428139: [toolsdb] Transaction History Length growing too much
T427187: ToolsDB disk space usage growing too fast

Event Timeline

fnegri created this task.Wed, Jun 3, 5:10 PM

Restricted Application added a project: cloud-services-team. · View Herald TranscriptWed, Jun 3, 5:10 PM

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

fnegri mentioned this in T427187: ToolsDB disk space usage growing too fast.Wed, Jun 3, 5:11 PM

JJMC89 added a project: Data-Services.Wed, Jun 3, 5:11 PM

JJMC89 moved this task from Backlog to Shared Storage on the Data-Services board.

JJMC89 removed a project: Data-Services.

fnegri mentioned this in T428139: [toolsdb] Transaction History Length growing too much.Fri, Jun 5, 11:34 AM

[toolsdb] Add db-level and user-level monitoringOpen, Needs TriagePublicActions

Description

Related Objects

Event Timeline

[toolsdb] Add db-level and user-level monitoring
Open, Needs TriagePublic
Actions