From 165d44ba735ed012644d024dcac3afe62f3c9aec Mon Sep 17 00:00:00 2001 From: Jonathan Irwin Date: Tue, 11 Aug 2026 15:45:39 -0400 Subject: [PATCH 1/3] docs(deepgram): note that the Engine port must stay 8055 Cerebrium detects that a Deepgram app is ready to serve traffic by watching the Engine's server port, so changing it in engine.toml makes requests arrive before the Engine can answer them and fail with 503 while the app starts up. Add a warning before the engine.toml block and call it out inline in both the engine.toml and api.toml snippets, since those get copied and edited. --- partner-services/deepgram.mdx | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/partner-services/deepgram.mdx b/partner-services/deepgram.mdx index d18f0cd3..0ba20bfd 100644 --- a/partner-services/deepgram.mdx +++ b/partner-services/deepgram.mdx @@ -60,6 +60,14 @@ cerebrium cp nova-3-general.en.streaming.123456.dg deepgram-models/nova-3-genera cerebrium cp engine.toml deepgram/engine.toml ``` + + Leave the Engine's `[server]` port set to **8055**. Cerebrium watches this port + to tell when the Engine has finished loading your models and the app is ready to + serve traffic. If you change it, requests are sent before the Engine can answer + them and fail with `503 Please try again later` while the app starts up. The + `driver_pool` URL in `api.toml` (step 5) must point at the same port. + + ```bash ### Keep in mind that all paths are in-container paths and do not need to exist @@ -84,7 +92,8 @@ server_url = ["https://license.deepgram.com"] ### The IP address to listen on. Since this is likely running in a Docker ### container, you will probably want to listen on all interfaces. host = "0.0.0.0" -### The port to listen on +### The port to listen on. On Cerebrium this must stay 8055 — it is how the +### platform detects that the Engine is ready to serve traffic. port = 8055 @@ -260,6 +269,8 @@ speak_streaming = true # or false ### ### Docker Compose and Podman Compose create a dedicated network that allows inter-container communication by app name. ### See [Networking in Compose](https://docs.docker.com/compose/networking/) for details. +### +### On Cerebrium, keep port 8055 here so it matches `[server]` in engine.toml. url = "https://0.0.0.0:8055/v2" ### Factor to increase the timeout by for each additional retry (for ### exponential backoff). From f7ebdb8b1b7e62198c9e9a1c5270110d5098813c Mon Sep 17 00:00:00 2001 From: jonoirwinrsa Date: Tue, 11 Aug 2026 19:46:02 +0000 Subject: [PATCH 2/3] Prettified Code! --- partner-services/deepgram.mdx | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/partner-services/deepgram.mdx b/partner-services/deepgram.mdx index 0ba20bfd..10b3d30e 100644 --- a/partner-services/deepgram.mdx +++ b/partner-services/deepgram.mdx @@ -61,11 +61,12 @@ cerebrium cp nova-3-general.en.streaming.123456.dg deepgram-models/nova-3-genera ``` - Leave the Engine's `[server]` port set to **8055**. Cerebrium watches this port - to tell when the Engine has finished loading your models and the app is ready to - serve traffic. If you change it, requests are sent before the Engine can answer - them and fail with `503 Please try again later` while the app starts up. The - `driver_pool` URL in `api.toml` (step 5) must point at the same port. + Leave the Engine's `[server]` port set to **8055**. Cerebrium watches this + port to tell when the Engine has finished loading your models and the app is + ready to serve traffic. If you change it, requests are sent before the Engine + can answer them and fail with `503 Please try again later` while the app + starts up. The `driver_pool` URL in `api.toml` (step 5) must point at the same + port. ```bash From a5453c624af394d8e4fd6bfdb3cc69acebce9378 Mon Sep 17 00:00:00 2001 From: Jonathan Irwin Date: Tue, 11 Aug 2026 15:49:41 -0400 Subject: [PATCH 3/3] docs(deepgram): lead with the queueing behaviour in the port note Say what keeping 8055 buys the user - startup requests are queued rather than returning 503 - rather than describing how readiness is detected. --- partner-services/deepgram.mdx | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/partner-services/deepgram.mdx b/partner-services/deepgram.mdx index 10b3d30e..4960b3c0 100644 --- a/partner-services/deepgram.mdx +++ b/partner-services/deepgram.mdx @@ -61,12 +61,10 @@ cerebrium cp nova-3-general.en.streaming.123456.dg deepgram-models/nova-3-genera ``` - Leave the Engine's `[server]` port set to **8055**. Cerebrium watches this - port to tell when the Engine has finished loading your models and the app is - ready to serve traffic. If you change it, requests are sent before the Engine - can answer them and fail with `503 Please try again later` while the app - starts up. The `driver_pool` URL in `api.toml` (step 5) must point at the same - port. + Keep the Engine's `[server]` port set to **8055**, and the `driver_pool` URL + in `api.toml` pointing at it. Cerebrium uses this port to tell when the Engine + is ready, so requests that arrive during startup are queued instead of failing + with `503 Please try again later`. ```bash @@ -93,8 +91,8 @@ server_url = ["https://license.deepgram.com"] ### The IP address to listen on. Since this is likely running in a Docker ### container, you will probably want to listen on all interfaces. host = "0.0.0.0" -### The port to listen on. On Cerebrium this must stay 8055 — it is how the -### platform detects that the Engine is ready to serve traffic. +### The port to listen on. On Cerebrium, keep 8055 so requests are queued while +### the Engine starts up. port = 8055