[SmartSwitch] Support maximum gnmi message size of 125k messages

[croos12]

Why I did it

At max DASH scale, a single gNMI call (CA-to-PA, ENI-level object config) can contain up to ~125,000 protobuf messages, which serializes to ~32 MB. sonic-gnmi currently uses gRPC's 4 MB default for MaxRecvMsgSize, so these calls are rejected.

How I did it

Setting	Before	After
gRPC MaxRecvMsgSize (server)	4 MB (gRPC default)	configurable via -max_recv_msg_size
gRPC MaxSendMsgSize (server)	math.MaxInt32 (default)	configurable via -max_send_msg_size

Defaults are unchanged; sonic-buildimage#26679 sets both to 32 MB for SmartSwitch.

How to verify it

Start telemetry with -max_recv_msg_size above 4 MB and send a gNMI message larger than 4 MB; confirm it is accepted.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[Copilot]

Pull request overview

This PR adds configurable gRPC max receive/send message sizes to the SONiC gNMI telemetry server, to better handle larger config/telemetry payloads at scale.

Changes:

• Added MaxRecvMsgSize / MaxSendMsgSize to TelemetryConfig.
• Introduced -max_recv_msg_size and -max_send_msg_size CLI flags.
• Applied the configured limits to the gRPC server via grpc.MaxRecvMsgSize / grpc.MaxSendMsgSize.

[vivekrnv]

@prsunny PFA

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[croos12]

/azpw run

[mssonicbld]

⚠️ Notice: /azpw run only runs failed jobs now. If you want to trigger a whole pipline run, please rebase your branch or close and reopen the PR.
💡 Tip: You can also use /azpw retry to retry failed jobs directly.

Retrying failed(or canceled) jobs...

[mssonicbld]

Retrying failed(or canceled) stages in build 1085425:

✅Stage Build:

• Job build: retried.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[hdwhdw]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[croos12]

/azpw run

[mssonicbld]

⚠️ Notice: /azpw run only runs failed jobs now. If you want to trigger a whole pipline run, please rebase your branch or close and reopen the PR.
💡 Tip: You can also use /azpw retry to retry failed jobs directly.

Retrying failed(or canceled) jobs...

[mssonicbld]

Retrying failed(or canceled) stages in build 1085623:

✅Stage BuildArm64:

• Job arm64 deb build: retried.

✅Stage BuildAmd64:

• Job amd64 deb build: retried.

✅Stage Build:

• Job amd64 deb build: retried.
• Job arm64 deb build: retried.
• Job build: retried.

[croos12]

/azpw retry

[mssonicbld]

Retrying failed(or canceled) jobs...

[mssonicbld]

Retrying failed(or canceled) stages in build 1085623:

✅Stage BuildAmd64:

• Job amd64 deb build: retried.

✅Stage BuildArm64:

• Job arm64 deb build: retried.

✅Stage Build:

• Job amd64 deb build: retried.
• Job build: retried.
• Job arm64 deb build: retried.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[hdwhdw]

@croos12 Just to double check do you mean 125K or 125M in title?

[croos12]

@croos12 Just to double check do you mean 125K or 125M in title?

It's 125k total messages per file max. In testing I found that will fit in 32MB for gnmi send/recv buffers

[hdwhdw]

Sorry this is still a bit confusing. I don't see any mentioning of the 125kb in the new code and isn't 125 kb < 4mb? What exactly are you changing to 125kb?

[croos12]

Sorry this is still a bit confusing. I don't see any mentioning of the 125kb in the new code and isn't 125 kb < 4mb? What exactly are you changing to 125kb?

Sorry these variables are kind of confusing. The "125k" in the title means 125,000 messages (count), not 125 kilobytes. The largest ENI-level object configuration in DASH requires up to 125,000 protobuf messages to be sent in a single gNMI call (CA-to-PA). At that scale, the serialized payload requires a maxMessage size of around ~32MB, which I am setting in my sonic-buildimage PR that requires this change: sonic-net/sonic-buildimage#26679

[hdwhdw]

Thanks for the clarification. It would be nice to include some background or pointers to background in the description.

Also you might want to merge latest master for the pipeline error.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).