Bjorn Leffler has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/541330 ) Change subject: Add Prometheus and Grafana monitoring stack to NFS Ganesha. ...................................................................... Add Prometheus and Grafana monitoring stack to NFS Ganesha. Measure, export, aggregate and display monitoring metrics for: 1. NFS requests, aggregated per {operation} and per {operation, export} - Request rates - Sent / received throughput rates in bytes/second. - Latency percentiles in ms as a histogram. - Request and Response sizing in bytes as a histogram. 2. Number of NFS worker threads in use. 3. Number of active NFS clients. 4. Metadata cache hit and miss rates, per {operation} and per {operation, export} 5. Rates of received and completed rpcs. 6. Number of RPCs in flight. 7. Per NFS client request and throughput rates. This change adds the first C++ file to NFS Ganesha. I tried to implement the monitoring in C only using the digitalocean C Prometheus client. This turned out to be a bottleneck that wasted 25% of all cpu and limited overall request throughput (https://github.com/digitalocean/prometheus-client-c/issues/59). This flatlined Ganesha requests/second performance. As there are no other C libraries, the monitoring code instead uses another C++ library, which doesn't have this performance problem. The C++ library is thread safe, so no locking is added in NFS Ganesha. By default, export metrics on TCP port 9587. Reuse the same port that is allocated to the nfs-ganesha exporter at https://github.com/prometheus/prometheus/wiki/Default-port-allocations Other changes: - Add a parameter to NFS core parameters to allow the user change the monitoring port. - As this change adds new library dependencies, guard it by the USE_MONITORING definition. - Add a new log component, COMPONENT_MONITORING - Add sample configuration file for Prometheus metrics aggregation. - Add sample configuration file for Grafana console. Signed-off-by: Bjorn Leffler <leffler(a)google.com&gt; Change-Id: I825cf2283f6e9e8c450cbf2923f92137fbee7e58 --- M src/CMakeLists.txt M src/MainNFSD/CMakeLists.txt M src/MainNFSD/nfs_main.c M src/config_samples/config.txt A src/config_samples/grafana.dashboard.json A src/config_samples/prometheus.rules.yml M src/include/config-h.in.cmake M src/include/gsh_config.h M src/include/log.h A src/include/monitoring.h M src/log/log_functions.c A src/monitoring/CMakeLists.txt A src/monitoring/monitoring.cc A src/monitoring/monitoring_internal.h M src/support/fridgethr.c M src/support/nfs_read_conf.c M src/support/server_stats.c 17 files changed, 4,228 insertions(+), 12 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/30/541330/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/541330 To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: I825cf2283f6e9e8c450cbf2923f92137fbee7e58 Gerrit-Change-Number: 541330 Gerrit-PatchSet: 1 Gerrit-Owner: Bjorn Leffler <leffler(a)google.com&gt; Gerrit-MessageType: newchange