Add instance age metric

[vmercierfr]

Objective

Add instance age metric

Why

Some workloads require raising alerts based on instance age.

At Qonto, we raise alerts when some instances (based on name patterns) run too long to prevent uncontrolled AWS costs.

How

• Add rds_instance_age_seconds
• Document how to add a new metric

Release plan

• Merge this PR

[vmercierfr]

#sre

[vmercierfr]

Reviewer's note: The rds_instance_uptime metric in rds-management was a little misleading, as it appears to be the time since the RDS instance was last start, when it is the creation date. rds_instance_age_seconds sounds more accurate.

[qfritz]

LGTM

Maybe if we have the need for it, future releases can include an rds_instance_uptime which will completely disambiguate rds_instance_age.

[dcupif]

lgtm 🚀

[vmercierfr]

I've added the docs/add-metric.md file to explain how to add a metric for future metrics.

[github-actions]

Package	Line Rate	Health
github.com/qonto/prometheus-rds-exporter/internal/app/cloudwatch	88%	✔
github.com/qonto/prometheus-rds-exporter/internal/app/ec2	88%	✔
github.com/qonto/prometheus-rds-exporter/internal/app/rds	79%	➖
github.com/qonto/prometheus-rds-exporter/internal/app/servicequotas	51%	❌
github.com/qonto/prometheus-rds-exporter/internal/app/unit	100%	✔
Summary	82% (523 / 640)	✔

Minimum allowed line rate is 60%