Handle bot launch error so the app won't crash #1146
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
mfahmialkautsar
force-pushed
the
main
branch
from
dbae2ec
to
52dfe5f
Compare
mfahmialkautsar
force-pushed
the
main
branch
from
52dfe5f
to
6cbc47d
Compare
|
Any prescribed approach on how to avoid this error? My app is also facing this issue when starting up cloudrun instances in GCP where multiple instances are being spawned around the same time which causes the 429s when setting the webhook. |
This issue happens because the Telegram webhook API is called every time the app starts, which is not suitable for serverless environments. I previously suggested a change to handle this by ignoring the error and allowing the app to continue running, assuming the webhook might have already been set in a previous call. Since the maintainers didn't accept the PR, I forked the repo and made the change myself. You can use my fork by replacing the source to be "nestjs-telegraf": "github:mfahmialkautsar/nestjs-telegraf#dist" in your package.json. Note that I haven't updated the fork since then, and it's more of a temporary workaround. Ideally, I think the best approach would be to create a dedicated API endpoint for registering the webhook, which you can call manually when needed. This way, the webhook setup doesn't happen on every app start, making it very efficient and reducing the chance of hitting rate limits. |
I am always annoyed when my app is deployed to Vercel which is serverless (so the app must be started at every function call), there's always a chance for the nestjs-telegraf to throw error API call when launching, such as:
2023-09-27T05:49:41.056Z 1c417191-68c4-4c4a-bcb5-55210c2e2845 ERROR Unhandled Promise Rejection {"errorType":"Runtime.UnhandledPromiseRejection","errorMessage":"Error: 429: Too Many Requests: retry after 1","reason":{"errorType":"Error","errorMessage":"429: Too Many Requests: retry after 1","code":429,"response":{"ok":false,"error_code":429,"description":"Too Many Requests: retry after 1","parameters":{"retry_after":1}},"on":{"method":"setWebhook","payload":{"url":"https://myapp-a3kd7o5r5-myusername.vercel.app/my-secret-path"}},"stack":["Error: 429: Too Many Requests: retry after 1"," at Telegram.callApi (/var/task/node_modules/.pnpm/telegraf@4.14.0/node_modules/telegraf/lib/core/network/client.js:294:19)"," at processTicksAndRejections (node:internal/process/task_queues:95:5)"," at Telegraf.launch (/var/task/node_modules/.pnpm/telegraf@4.14.0/node_modules/telegraf/lib/telegraf.js:198:9)"]},"promise":{},"stack":["Runtime.UnhandledPromiseRejection: Error: 429: Too Many Requests: retry after 1"," at process.<anonymous> (file:///var/runtime/index.mjs:1250:17)"," at process.emit (node:events:526:35)"," at process.emit (/var/task/___vc/__sourcemap_support.js:602:21)"," at emit (node:internal/process/promises:149:20)"," at processPromiseRejections (node:internal/process/promises:283:27)"," at processTicksAndRejections (node:internal/process/task_queues:96:32)"]} [ERROR] [1695793781058] LAMBDA_RUNTIME Failed to post handler success response. Http response code: 400. RequestId: 7b50cdab-42c0-42ee-b707-e2392b984d1b Error: Runtime exited with error: exit status 128 Runtime.ExitErrorSo I think it would be better to ignore the error and let the app run without nestjs-telegraf module. Because it's just a webhook and if my app has another API that's called from other client, it won't need the Telegram webhook. In serverless case, it wouldn't matter for Telegram webhook since it will retry to call the endpoint again.