Philipp Hossner

CPU load high when backend server with healthcheck down

35

### Detailed Description of the Problem Hello everyone, I have a high-load (~10k reqs/s) HaProxy 2.6.1 instance with HTTP mode backends with L4 health checks. Everything works fine until a...

When using the --coverage option with a source file with a \_\_future\_\_ import statement at the top the following error occurrs: `SyntaxError: from __future__ imports must occur at the beginning...

Generic version of src_inc_gpc with key parameter

2

### Your Feature Request Hi, the fetch method `src_inc_gpc` can be used to put an entry for the source IP address of the current request into a stick table if...

Large NIC ring buffer sizes cause HAProxy v2.7.10 to crash under load

2

### Detailed Description of the Problem Hi, I tried to apply some NIC optimizations to all physical interfaces: ``` # systemd-networkd config section CombinedChannels=64 RxBufferSize=2047 TxBufferSize=2047 UseAdaptiveRxCoalesce=true UseAdaptiveTxCoalesce=true ``` The...

Volume directoy on host quickly disappears after mount

6

Hi, the volume on my DiskStation (2x 920+ in HA) is provisioned correctly and attached as block device (/dev/sdd in my case) on the host system. But it is expected...

I finally got around to implement the feature. The changes were bigger than expected. Sorry about that in advance. The `command` module now has end-to-end tests. For that to work...

Pseudonymization of common_name

4

Hi, since client metrics endanger privacy of the clients but are useful for debugging I propose the following feature: Each common_name is replaced with a random string of fixed length...

Add "retention" feature allowing idle metrics to expire

9

As proposed in #20 I implemented a way to auto remove idle metrics at runtime without needing to restart. Example: ``` name message_foo_counter type counter desc The total number of...

When having a lot of label combinations for a metric but only the metrics with large values are of interest (e.g. a `url` label) it might make sense to restrict...

Hi, the exporter might generate metric names with hyphens which causes prometheus scrape to crash with `invalid metric type "0 gauge"`. Example: ``` # HELP clickhouse_block_write_merges_rbd0 Number of BlockWriteMerges_rbd0 async...