gRPC

Last updated Jul 24, 2026

Connect AnyLog as a gRPC client to receive data streams from gRPC servers and map them to local database tables.

---

gRPC is Google’s open-source RPC framework — efficient, language-agnostic, and designed for high-throughput streaming. AnyLog connects as a gRPC client, receives data streams from a gRPC server, and maps them to a local database using policies.

---

Setup

1. Install grpcio-tools

python3 -m pip install --upgrade grpcio-tools

2. Create a proto file

The proto file defines the service, RPC methods, and message types. Example dummy.proto:

syntax = "proto3";

package mygrpc;

service SerializeService {
  rpc GetSampleData (Empty) returns (SampleDataResponse);
}

message Empty {}

message SampleDataResponse {
  repeated string serialized_data = 1;
}

3. Compile the proto file

Run from the same directory as the .proto file:

python3 -m grpc_tools.protoc -I. --python_out=. --grpc_python_out=. dummy.proto

This generates two files (e.g. dummy_pb2.py and dummy_pb2_grpc.py) that AnyLog uses to communicate with the server.

---

Starting a gRPC client

<run grpc client where
  name    = [unique name] and
  ip      = [server IP] and
  port    = [server port] and
  grpc_dir = [path to proto/compiled files] and
  proto   = [proto file name, no extension] and
  function = [RPC method name] and
  request = [request message type] and
  response = [response message type] and
  service  = [service name] and
  policy   = [mapping policy ID] and
  value    = (key = value and ...) and
  debug   = [true/false] and
  limit   = [max events] and
  dbms    = [target database] and
  table   = [target table] and
  ingest  = [true/false] and
  invoke  = [true/false]>

Parameter reference

Parameter	Required	Description
name	✅	Unique ID for this gRPC connection
ip	✅	gRPC server IP
port	✅	gRPC server port
grpc_dir	✅	Directory containing the compiled proto files
proto	✅	Proto base filename (without extension)
function	✅	RPC method to call on the server
request	✅	Request message type name from the proto
response	✅	Response message type name from the proto
service	✅	Service name from the proto
policy	—	Mapping policy ID for schema mapping
value	—	Key-value pairs sent in the request message (e.g. Filter = system and Type = 5.int)
debug	—	true prints received data to the CLI (default: false)
limit	—	Stop after N events
dbms	—	Target database (if not provided by policy)
table	—	Target table (if not provided by policy)
ingest	—	false disables database ingestion — useful for testing (default: true)
add_info	—	Append metadata to each row — see Options for add_info below
invoke	—	Whether to invoke immediately when called by another AnyLog process

Examples

<run grpc client where
  name = kubearmor and ip = 127.0.0.1 and port = 50051 and
  grpc_dir = /app/AnyLog-Network/proto/kubearmor and
  proto = kubearmor and function = WatchLogs and
  request = RequestMessage and response = Log and
  service = LogService and value = (Filter = policy) and
  debug = true and limit = 2 and ingest = false>

<run grpc client where
  name = kubearmor and ip = 127.0.0.1 and port = 50051 and
  grpc_dir = /app/AnyLog-Network/proto/kubearmor and
  proto = kubearmor and function = HealthCheck and
  request = NonceMessage and response = ReplyMessage and
  service = LogService and value = (nonce = 10.int) and
  debug = true and limit = 1 and ingest = false>

Options for add_info

When add_info is included in the run grpc client command, the specified keys are added as extra fields to the JSON data retrieved from the server. add_info can be repeated — once per key you want added:

Key	Value added to the JSON
proto	The name of the proto file
request	The name of the request message in the proto file
conn	The IP and port used

<run grpc client where name=kubearmor and ip = 127.0.0.1 and port = 50051 and
  grpc_dir = /app/AnyLog-Network/proto/kubearmor and
  proto = kubearmor and function = WatchLogs and policy = kubearmor-system-policy and
  request = RequestMessage and response = Log and service = LogService and
  value = (Filter = all) and debug = false and limit = 10000 and ingest = false and
  add_info = conn and add_info = proto and add_info = request>

---

Monitor and manage

# List all active gRPC clients
get grpc clients

Example output:

ID        Connection       Proto     Request Message  Policy ID                Timeouts  Data Msg
---------|----------------|---------|----------------|------------------------|--------|--------|
health   |10.0.0.251:32769|kubearmor|NonceMessage    |                        |       0|    1254|
kubearmor|10.0.0.251:32769|kubearmor|RequestMessage  |kubearmor-system-policy |       0|       0|

# List services offered by the gRPC server (requires server reflection)
get grpc services where conn = [ip:port]

# Stop a specific client
exit grpc [name]

# Stop all clients
exit grpc all

---

gRPC for video inference

When used with video streaming, the gRPC client connects to a YOLOv5 (or similar) inference server. See Video Streaming for the full configuration including video connect and run video stream.

node_configs.env