[purger] Ensure docker restart timeout is large enough

[konradkonrad]

Problem Description

• docker stop / restart has a default timeout of 10s before it issues a kill -9 to the process
• purger.py currently uses a timeout of 30s
• state_groups_state size explosion seems related to orphaned state groups
• cleanup tools i.e. https://github.com/erikjohnston/synapse-find-unreferenced-state-groups urge you to "properly shutdown" synapse to avoid problems around "unsealed" state groups
• manual docker stop -t 120 <synapse_container> did sometimes not terminate without the kill -9 for me
• so there is the possibility that container restarts add to the state explosion problem (even though it is only done on whitelist changes).