This issue does seem specific to this Node-RED-Heroku repo that I used to install Node-RED as a Heroku element.
I am considering completely dumping this app but thought I'd be able to get some guidance for salvaging this in the meantime.
tl;dr: in the last 7 days, the Node-RED heroku apps have been throwing a timeout err error: no pg_hba.conf entry for host ""...
error even though I have made no changes to the underlying databases or any of the code in the Node-RED app that makes calls to Postgres
Here are the full logs when attempting to restart the heroku dynos:
2021-03-16T05:40:15.998255+00:00 heroku[web.1]: Starting process with command `node --max-old-space-size=384 node_modules/node-red/red.js --settings ./settings.js`
2021-03-16T05:40:19.832866+00:00 app[web.1]: pgUrl postgres://<redacted>
2021-03-16T05:40:19.834452+00:00 app[web.1]: create pg tables
2021-03-16T05:40:20.241122+00:00 app[web.1]: (node:4) UnhandledPromiseRejectionWarning: error: no pg_hba.conf entry for host "<redacted>", user "<redacted>", database "<redacted>", SSL off
//truncated...
2021-03-16T05:40:20.302015+00:00 app[web.1]: TIMEOUT: getSettings
2021-03-16T05:40:20.397532+00:00 heroku[web.1]: Process exited with status 0
2021-03-16T05:40:20.565037+00:00 heroku[web.1]: State changed from starting to crashed
This behavior is repeatable across all 4 of my Node-RED heroku apps any time I attempt to restart the dynos.
From what I can tell, it seems there's the getSettings
function in pgstorage.js
is experiencing timeout and that's now causing the app to crash.
From what I am seeing, the "fix" seems to be changing a property value in a config, I'm still figuring out how to do that in Heroku context because the machines themselves are stateless