Request to submit a batch of log records to the server. More...
#include <log_records_protocol.hpp>
Collaboration diagram for submit_log_records_request:
Public Member Functions | |
| std::vector< std::byte > | serialize () const |
| Serialize request to bytes. | |
Static Public Member Functions | |
| static std::expected< submit_log_records_request, ores::utility::serialization::error_code > | deserialize (std::span< const std::byte > data) |
| Deserialize request from bytes. | |
Public Attributes | |
| std::vector< domain::log_record > | records |
| The batch of log records to submit. | |
Detailed Description
Request to submit a batch of log records to the server.
This is a fire-and-forget message. The client does not wait for a response and continues processing immediately. The server stores the records in the telemetry database for analysis and correlation.
Records should be batched for efficiency. Typical batch sizes are 50-100 records, or send when a flush interval (e.g., 5 seconds) has elapsed.
Member Function Documentation
◆ serialize()
| std::vector< std::byte > serialize | ( | ) | const |
Serialize request to bytes.
Format:
- 4 bytes: record count (uint32) For each record:
- 8 bytes: timestamp (milliseconds since epoch, uint64)
- 1 byte: severity level (uint8)
- 2 bytes: body length
- N bytes: body (UTF-8)
- 2 bytes: logger_name length
- N bytes: logger_name (UTF-8)
- 1 byte: has_trace (boolean)
- 32 bytes: trace_id hex string (if has_trace)
- 1 byte: has_span (boolean)
- 16 bytes: span_id hex string (if has_span)
- 2 bytes: service_name length
- N bytes: service_name (UTF-8, from resource if present)
