chore: update js-ceramic healthcheck timing

[3benbox]

• lower initial_delay_seconds so that healthchecks are run faster and pods are marked up earlier
• lower period_seconds so that there are less healthchecks (already enough load on the system!)
• increase timeout_seconds as the js-ceramic health check makes its own external calls to "ipfs" which where longer than this setting

[dav1do]

This seems aok and I like reducing the initial_delay_seconds but find it hard to believe the timeouts and period were too low (it makes me sad).

We could consider using different values for startup and liveness like this if we're still seeing odd behavior, or possibly adjust the failureThreshold.