Logs and metrics read from the Loggregator V2 Firehose over HTTPS (the `log-stream` endpoint) drop under load

Products

• VMware Tanzu Application Service for VMs

Issue/Introduction

When reading logs from the V2 Firehose using the `https://log-stream.SYSTEM_DOMAIN` endpoint, you experience heavy loss under load due to performance issues in the RLP Gateway server and client.

Resolution

There is currently no workaround for this issue and the team is investigating a fix.

The following logs or metrics read via the following methods do not experience this issue:

• The V1 Firehose (`https://doppler.SYSTEM_DOMAIN` endpoint).
• Syslog drains per-app or the aggregate drains in Tanzu Application Service (TAS) for VMs 2.8+.
• The V2 Firehose via the internal gRPC endpoint.