-
Spike
-
Resolution: Done
-
Major
-
None
-
None
-
0
-
False
-
None
-
False
-
Yes
-
Sprint 23, MGDAPI - Sprint 24
WHAT
As RHOAM on Developer Sandbox moves closer to production, we need a way to track multitenant specific metrics. First we need to decide which metrics we want to expose. Below is a list of potentially useful metrics to track; note that the list is not exhaustive and not everything listed may be useful. Additionally, some of these metrics are already created in integreatly-operator
Potentially useful metrics:
- Number of reconciled tenants
- Number of failed tenants and why they failed (i.e. tenant CR created in wrong namespace, user for tenant CR doesn't exist, tenant CR already exists, etc.)
- Number of user signups
- Number of incomplete user signups and why they failed (i.e. end user started account creation but didn't verify email, SMS, etc.)
- Total resource consumption by tenant (API requests, CPU, memory, etc.)
- Total resource consumption from all tenants (API request, CPU, memory, etc.)
HOW
Implement the Requirement metrics, Prom Queries, Alerts, Grafana Dashboards etc
NOTE: There are some metrics we can use out of the box from Sandbox Operators
An interim solution might be a simple script to retrieve some of the above data.
DONE
- Determined which metrics we want to expose
- Implement the required metrics, Prom Queries, Alerts, and Dashboards where applicable