You may have noticed that when we run a test, `nb`

tells us about logging.
For example, we may see a line like this:

```
Logging to logs/scenario_20200429_160016_752.log
```

Let's clear out the logs and re-run the test.

```
rm -rf logs
./nb start driver=cql workload=cql-keyvalue tags=phase:main cycles=100k cyclerate=5000 threads=50 --progress console:2s
```

If you look at the log files, they can be a bit overwhelming because they contain a lot of trace information.
We are interested in performance information, so we can filter the log file with this command.

```
cat logs/scenario_* | grep 'type=TIMER' | head -n 1
```

This line shows performance metrics.
Let's look at an example to see how to interpret these lines.

```
2020-04-29 16:57:11,414 INFO [main] i.n.e.c.ScenarioResult [Slf4jReporter.java:374] type=TIMER, name=cql-keyvalue.bind, count=100000, min=0.392, max=16811.007, mean=6.76477675, stddev=173.637642943727, median=1.479, p75=2.314, p95=8.652, p98=17.374, p99=35.603, p999=482.063, mean_rate=4323.519012492167, m1=3726.7032056089, m5=3527.812026532018, m15=3490.6484917755765, rate_unit=events/second, duration_unit=microseconds
```

`2020-04-29 16:57:11`

indicates when the sample measurement occurred
`414 INFO [main] i.n.e.c.ScenarioResult [Slf4jReporter.java:374] type=TIMER`

tells where, in the `nb`

code, the measurement came from
`name=cql-keyvalue.bind`

indicates the name of the wokload file
`count=100000`

is the number of operations in this sample
`min=0.392`

is the time measurement of the shortest operation (see the `duration_unit`

at the bottom of the list for the units)
`max=16811.007`

is the time measurement of the longest operation
`mean=6.76477675`

is the mean (average) operation time
`stddev=173.637642943727`

is the standard deviation from the mean
`median=1.479`

is the median (middle) operation time
`p75=2.314`

is the 75th percentile measurement (75% of operations took this amount of time or less)
`p95=8.652`

is the 95th percentile measurement
`p98=17.374`

is the 98th percentile measurement
`p99=35.603`

is the 99th percentile measurement
`p999=482.063`

is the 99.9th percentile measurement
`mean_rate=4323.519012492167`

is the number of operations per time period (see the `rate_unit`

near the bottom of the list for the units)
`m1=3726.7032056089`

is the mean rate measured during 1 minute
`m5=3527.812026532018`

is the mean rate measured during 5 minutes
`m15=3490.6484917755765`

is the mean rate measured during 15 minutes
`rate_unit=events/second`

indicates the time unit for rate measurements
`duration_unit=microseconds`

indicates the time unit for timing measurements

There are many ways to process the log file output and many helpful tools.
Check these out here.

## You now have a sense of what NoSQLBench can do for you. But, we've only scratched the surface. Learn more here!