# Known Issues ## Frontend sandbox deployments uses an outdated engine image When deploying a single repository, the counterpart image should be always pulled from the latest `master` commit. This will ensure the review site is using the most up-to-date version of the stack. ## Cron jobs TBD. - Cron jobs not running currently on sandboxes - Cron jobs need to run on a shared pod that would be updateable ## Settings management Settings are prone to suffer inconsistencies upon deployment updates. This can be achieved by: - Making `Core\Config` to throw an `Exception` when a key doesn't exist (or is null). This should be implemented taking in consideration that old code should still work. - Enforcing the developers to keep track of their new configurations in `settings.example.php` - Having a base `settings.php` with sensitive defaults (except for keys, etc), then having an "override" file that will be used on local envs, sandboxes and even production. ## Missing posts and media uploads Sometimes posts (with or without attachments) are not shown in any feed, but they can be accessed by browsing to `https://sandboxsite/newsfeed/<guid>`. To be investigated. Probably related to ElasticSearch health. Try reproducing when E2E tests are running. ## ElasticSearch freeze Sometimes it takes around 30s to receive a response from ElasticSearch-based endpoints (feeds, pageview beacons, etc). To be investigated. Probably related to ElasticSearch health. Try reproducing when E2E tests are running. ## Pro subdomains TBD. Ideally we would have `*.<sandbox_domain>.minds.io` frontend on all sandbox deployments. ## OnChain tokens / Rinkeby support Sandboxes should have all the keys set up as Helm secrets to allow OnChain transactions on smoke and E2E tests.