github.com/m3db/m3@v1.5.0/src/cmd/tools/annotation_checker/main/README.md (about)

     1  # annotation_checker
     2  
     3  `annotation_checker` is a utility that reads TSDB file sets and analyzes timeseries for missing initial annotations or annotation rewrites.
     4  
     5  By default, this tools output the list of series IDs that have no annotation with the first datapoint in the block. If `--annotation-rewritten` flag is used, the output will contain series IDs of metrics with rewritten annotations instead.
     6  
     7  # Usage
     8  ```
     9  $ make annotation_checker
    10  $ ./bin/annotation_checker
    11  Usage: annotation_checker [-R] [-b value] [-n value] [-p value] [-s value] [-t value] [-v value] [parameters ...]
    12   -b, --block-start=value
    13                      Block Start Time [in nsec]
    14   -n, --namespace=value
    15                      Namespace [e.g. metrics]
    16   -p, --path-prefix=value
    17                      Path prefix [e.g. /var/lib/m3db]
    18   -R, --annotation-rewritten
    19                      Filters metrics with annotation rewrites
    20   -s, --shard=value  Shard number, or -1 for all shards in the directory
    21   -t, --fileset-type=value
    22                      flush|snapshot
    23   -v, --volume=value
    24                      Volume number
    25  
    26  # example usage
    27  # annotation_chekcer -p /var/lib/m3db -n metrics -s 451 -b 1480960800000000000
    28  ```
    29  
    30  # TBH
    31  - The tool outputs the identifiers to `stdout`, remember to redirect as desired.
    32  - The code currently assumes the data layout under the hood is `<path-prefix>/data/<namespace>/<shard>/...<block-start>-[index|...].db`. If this is not the file structure under the hood, replicate it to use this tool. Remember to copy checkpoint files along with each index file.