improved README.md

Author: markuslf

disk-io/README.md

@@ -1,5 +1,12 @@
 # Check "disk-io" - Overview
 
+Checks the disk throughput over a certain amount of time. Warns only if any of your disks is above a certain throughput threshold within the last n checks (default: 5).
+
+Hints and Recommendations:
+* `--count=5` (the default) while checking every minute means that the check reports a warning if any of your disks was above a threshold in the last 5 minutes.
+* Check uses a SQLite database in `/tmp` to store its historic data.
+* If you are wondering about `dm-0`, `dm-1` etc.: It's part of the "device mapper" in the kernel, used by LVM. Use `dmsetup ls` to see what is behind it.
+
 We recommend to run this check every minute.
 
 
@@ -10,23 +17,30 @@ Requirements:
 
 ```bash
 ./disk-io
+./disk-io --ignore sr0 --ignore dm-1 --warning 100 --critical 200 --count 15
 ./disk-io --help
 ```
 
 
 # States
 
-tbd
+* OK if disk throughput of all disks is below the thresholds within the last `--count` checks.
+* Otherwise CRIT or WARN.
 
 
 # Perfdata
 
-tbd
+Per disk:
+
+* read_bytes_per_second: number of bytes read
+* read_count_per_second: number of reads
+* write_bytes_per_second: number of bytes written
+* write_count_per_second: number of writes
 
 
 # Known Issues and Limitations
 
-tbd
+There is no latency measurement due to the fact that the current `psutil` version (5.2) does not expose latency stats.
 
 
 # Credits, License