When sending metrics for 10,000 tables and indexes to PPEM via otlphttp, you may encounter the following error in pgpro-otel-collector logs:

{
    "level": "error",
    "ts": "2025-09-05T17:44:37.351+0300",
    "msg": "Exporting failed. Dropping data.",
    "resource": {
        "service.instance.id": "62cc1e9c-a53f-423e-9c6f-41b1f6a0872a",
        "service.name": "pgpro-otel-collector",
        "service.version": "v0.4.0"
    },
    "otelcol.component.id": "otlphttp/ppem_metrics",
    "otelcol.component.kind": "exporter",
    "otelcol.signal": "metrics",
    "error": "not retryable error: Permanent error: rpc error: code = Unknown desc = error exporting items, request to http://192.168.21.147:80/v1/metrics responded with HTTP Status Code 500, Message=, Details=[]",
    "dropped_items": 393313
}

This may indicate that PPEM cannot process the metrics batch within a single request. Use the batch processor to split the request to PPEM into smaller parts. Make sure to specify the send_batch_max_size parameter:

processors:
  batch/metrics:
    send_batch_size: 8192
    send_batch_max_size: 8192
    timeout: 10s
    ...

This parameter splits large requests into batches of 8,192 metrics each. After setting this parameter, ensure the sending queue does not overflow over time:

curl 127.0.0.1:8888/metrics | grep queue_size

The default maximum queue size is 1,000 elements. The more elements in the queue, the more RAM pgpro-otel-collector requires.

If the queue is full, the following error appears in the collector logs:

{
    "level": "warn",
    "ts": "2025-09-22T15:24:45.161+0300",
    "msg": "Sender failed",
    "resource": {
        "service.instance.id": "8859d053-3485-4a61-bc28-8d839bc9e20f",
        "service.name": "pgpro-otel-collector",
        "service.version": "v0.4.0"
    },
    "otelcol.component.id": "batch/metrics",
    "otelcol.component.kind": "processor",
    "otelcol.pipeline.id": "metrics",
    "otelcol.signal": "metrics",
    "error": "sending queue is full"
}

This means data is being dropped before being sent to PPEM. You need to reduce the number of metrics sent to PPEM or increase the queue size. Example queue configuration:

exporters:
  otlphttp/ppem_metrics:
    sending_queue:
      enabled: true
      queue_size: 1000
      ...

Additionally, you can increase batch/metrics.timeout and the global collection interval to collect metrics less frequently overall.

Retransmission of failed batches is enabled by default. Example configuration:

exporters:
  otlphttp/ppem_metrics:
    retry_on_failure:
      initial_interval: 5s
      max_interval: 30s
      max_elapsed_time: 300s
      multiplier: 1.5
      ...

The example shows default values. The complete set of queue and retry configuration options is available in the OpenTelemetry documentation.

The following error is related to the timeout specified on the PPEM side:

{
    "level": "info",
    "ts": "2025-09-05T18:20:28.465+0300",
    "msg": "Exporting failed. Will retry the request after interval.",
    "resource": {
        "service.instance.id": "b225f72c-f753-4816-85ef-57a982a0392c",
        "service.name": "pgpro-otel-collector",
        "service.version": "v0.4.0"
    },
    "otelcol.component.id": "otlphttp/ppem_metrics",
    "otelcol.component.kind": "exporter",
    "otelcol.signal": "metrics",
    "error": "failed to make an HTTP request: Post \"http://192.168.21.147:80/v1/metrics\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)",
    "interval": "4.905832986s"
}

On the PPEM side, increase the timeout as follows:

  http:
    server:
      timeout: 5m # Specify your value

Another error may be related to the timeout specified on the pgpro-otel-collector side, typically for logs and metrics:

{
    "level": "error",
    "ts": 1752593436.7439601,
    "msg": "Exporting failed. Dropping data.",
    "kind": "exporter",
    "data_type": "logs",
    "name": "otlphttp/ppem_logs",
    "error": "no more retries left: failed to make an HTTP request: Post \"http://192.168.22.114:80/v1/logs\": context deadline exceeded (Client.Timeout exceeded while awaiting headers)",
    "dropped_items": 1
}

Increase the timeout by specifying your values on the pgpro-otel-collector side:

exporters:
  otlphttp/elastic_logs:
    timeout: 1m
    ...
  otlphttp/ppem_logs:
    timeout: 1m
    ...
  otlphttp/ppem_metrics:
    timeout: 1m
    ...

Both errors may indicate that PPEM or pgpro-otel-collector lacks sufficient CPU resources to function properly with the specified plugin settings.