The issue:
When using the health check endpoint, The Edge Proxy is not currently able to withstand faults on the Flagsmith API due to the health check returning unhealthy if the environment data has not been updated in a defined period of time (as per the logic here).
Additional Considerations:
Since (currently) each edge proxy instance maintains its own cache of the environment data, we ideally want to know when a single instance no longer has an updated cache of the environment, but not when all instances are unable to update (because then it's likely that the Flagsmith API is down and we want the Edge Proxy to add as a cache). This may not be possible without a centralised cache.
The issue:
When using the health check endpoint, The Edge Proxy is not currently able to withstand faults on the Flagsmith API due to the health check returning unhealthy if the environment data has not been updated in a defined period of time (as per the logic here).
Additional Considerations:
Since (currently) each edge proxy instance maintains its own cache of the environment data, we ideally want to know when a single instance no longer has an updated cache of the environment, but not when all instances are unable to update (because then it's likely that the Flagsmith API is down and we want the Edge Proxy to add as a cache). This may not be possible without a centralised cache.