diff options
Diffstat (limited to 'doc/STATS.md')
| -rw-r--r-- | doc/STATS.md | 33 |
1 files changed, 33 insertions, 0 deletions
diff --git a/doc/STATS.md b/doc/STATS.md new file mode 100644 index 0000000..385d41e --- /dev/null +++ b/doc/STATS.md @@ -0,0 +1,33 @@ +Stats available through Management Server +========================================= + +Interface +--------- + +The management server makes statistics about the inputs available through a ZMQ request/reply socket. + +The `show_dabmux_stats.py` illustrates how to access this information. + +Meaning of values +----------------- + +`max` and `min` indicate input buffer fullness in bytes. + +`under` and `over` count the number of buffer underruns and overruns. + +`audio L` and `audio R` show the maximum audio level in dBFS over the last 500ms. + +`peak L` and `audio R` show the max audio level in dBFS over the last 5 minutes. + +The audio levels are measured in the audio encoder and carried in the EDI +`ODRa` TAG, or in the ZMQ metadata. Otherwise ODR-DabMux would have to decode +all audio contributions to measure the audio level. + +`State` is either NoData, Unstable, Silence, Streaming. +Unstable means that underruns or overruns have occurred in the previous 30 minutes. +Silence means the stream is working, but audio levels are always below -50dBFS. + +`version` and `uptime` are fields directly coming from the contribution source, +and are only supported for the EDI input. These are carried over EDI using custom +TAG `ODRv` (see function `parse_odr_version_data` in `lib/edi/common.cpp`). + |