aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorVratko Polak <vrpolak@cisco.com>2019-01-04 16:06:27 +0100
committerTibor Frank <tifrank@cisco.com>2019-02-18 07:38:37 +0000
commit55bcaeedebeba20cbdb48a0e6ddcc2c9f225697a (patch)
tree4978bc51925e3aa9d11ed669d668ca6c1991b0df
parenta1314a4934ba35e737b171767ba88b5e735541a0 (diff)
Doc: Methodology update for /new/ trending
Change-Id: Ibaf4594589cb7bcf1ad236c99f03a9f18b579a74 Signed-off-by: Vratko Polak <vrpolak@cisco.com> (cherry picked from commit 1aeacc2c1efb0ce39c2d0789ec04a694a0943ba7)
-rw-r--r--docs/cpta/methodology/index.rst122
1 files changed, 68 insertions, 54 deletions
diff --git a/docs/cpta/methodology/index.rst b/docs/cpta/methodology/index.rst
index c58d05fba9..7d7604bee8 100644
--- a/docs/cpta/methodology/index.rst
+++ b/docs/cpta/methodology/index.rst
@@ -6,30 +6,24 @@ Trending Methodology
Overview
--------
-// Below paragraph needs to be updated.
This document describes a high-level design of a system for continuous
performance measuring, trending and change detection for FD.io VPP SW
-data plane. It builds upon the existing FD.io CSIT framework with
-extensions to its throughput testing methodology, CSIT data analytics
-engine (PAL – Presentation-and-Analytics-Layer) and associated Jenkins
-jobs definitions.
-
-// Below paragraph needs to be updated.
-Proposed design replaces existing CSIT performance trending jobs and
-tests with new Performance Trending (PT) CSIT module and separate
+data plane (and other performance tests run within CSIT sub-project).
+
+There is a Performance Trending (PT) CSIT module, and a separate
Performance Analysis (PA) module ingesting results from PT and
analysing, detecting and reporting any performance anomalies using
-historical trending data and statistical metrics. PA does also produce
-trending dashboard and graphs with summary and drill-down views across
-all specified tests that can be reviewed and inspected regularly by
-FD.io developers and users community.
+historical data and statistical metrics. PA does also produce
+trending dashboard, list of failed tests and graphs with summary and
+drill-down views across all specified tests that can be reviewed and
+inspected regularly by FD.io developers and users community.
Performance Tests
-----------------
-Performance trending is relies on Maximum Receive Rate
-(MRR) tests. MRR tests measure the packet forwarding rate under the
-maximum load offered by traffic generator over a set trial duration,
+Performance trending relies on Maximum Receive Rate (MRR) tests.
+MRR tests measure the packet forwarding rate, in multiple trials of set
+duration, under the maximum load offered by traffic generator
regardless of packet loss. Maximum load for specified Ethernet frame
size is set to the bi-directional link rate.
@@ -70,6 +64,15 @@ using normal distribution. While sometimes the samples within a group
are far from being distributed normally, currently we do not have a
better tractable model.
+Here, "sample" should be the result of single trial measurement,
+with group boundaries set only at test run granularity.
+But in order to avoid detecting causes unrelated to VPP performance,
+the default presentation (without /new/ in URL)
+takes average of all trials within the run as the sample.
+Effectively, this acts as a single trial with aggregate duration.
+
+Performance graphs always show the run average (not all trial results).
+
The group boundaries are selected based on `Minimum Description Length`_.
Minimum Description Length
@@ -101,20 +104,16 @@ on the number of samples and stdev.
A somehow ambiguous part is in choosing which encoding
is used for group size, average and stdev.
-Diferent encodings cause different biases to large or small values.
+Different encodings cause different biases to large or small values.
In our implementation we have chosen probability density
corresponding to uniform distribution (from zero to maximal sample value)
for stdev and average of the first group,
but for averages of subsequent groups we have chosen a distribution
-which disourages deliminating groups with averages close together.
-
-// Below paragraph needs to be updated.
-One part of our implementation which is not precise enough
-is handling of measurement precision.
-The minimal difference in MRR values is currently 0.1 pps
-(the difference of one packet over 10 second trial),
-but the code assumes the precision is 1.0.
-Also, all the calculations assume 1.0 is totally negligible,
+which disourages delimiting groups with averages close together.
+
+Our implementation assumes that measurement precision is 1.0 pps.
+Thus it is slightly wrong for trial durations other than 1.0 seconds.
+Also, all the calculations assume 1.0 pps is totally negligible,
compared to stdev value.
The group selection algorithm currently has no parameters,
@@ -138,6 +137,9 @@ Once the trend data is divided into groups, each group has its population averag
The start of the following group is marked as a regression (or progression)
if the new group's average is lower (higher) then the previous group's.
+In the text below, "average at time <t>", shorthand "AVG[t]"
+means "the group average of the group the sample at time <t> belongs to".
+
Trend Compliance
````````````````
@@ -148,7 +150,6 @@ ago, AVG[last - 1week] and to the maximum of trend values over last
quarter except last week, max(AVG[last - 3mths]..ANV[last - 1week]),
respectively. This results in following trend compliance calculations:
-// Below table needs to be updated.
+-------------------------+---------------------------------+-----------+-------------------------------------------+
| Trend Compliance Metric | Trend Change Formula | Value | Reference |
+=========================+=================================+===========+===========================================+
@@ -166,30 +167,39 @@ Performance Dashboard
Dashboard tables list a summary of per test-case VPP MRR performance
trend and trend compliance metrics and detected number of anomalies.
-Separate tables are generated for tested VPP worker-thread-core
-combinations (1t1c, 2t2c, 4t4c). Test case names are linked to
-respective trending graphs for ease of navigation thru the test data.
+Separate tables are generated for each testbed and each tested number of
+physical cores for VPP workers (1c, 2c, 4c). Test case names are linked to
+respective trending graphs for ease of navigation through the test data.
+
+Failed tests
+````````````
+
+The Failed tests tables list the tests which failed over the specified seven-
+day period together with the number of fails over the period and last failure
+details - Time, VPP-Build-Id and CSIT-Job-Build-Id.
+
+Separate tables are generated for each testbed. Test case names are linked to
+respective trending graphs for ease of navigation through the test data.
Trendline Graphs
````````````````
-Trendline graphs show per test case measured MRR throughput values with
-associated gruop averages. The graphs are constructed as follows:
+Trendline graphs show measured per run averages of MRR values,
+group average values, and detected anomalies.
+The graphs are constructed as follows:
-- X-axis represents performance trend job build Id (csit-vpp-perf-mrr-
- daily-master-build).
-- Y-axis represents MRR throughput in Mpps.
+- X-axis represents the date in the format MMDD.
+- Y-axis represents run-average MRR value in Mpps.
- Markers to indicate anomaly classification:
- Regression - red circle.
- Progression - green circle.
-- The line shows average of each group.
+- The line shows average MRR value of each group.
In addition the graphs show dynamic labels while hovering over graph
-data points, representing (trend job build Id, MRR value) and the actual
-vpp build number (b<XXX>) tested.
-
+data points, presenting the CSIT build date, measured MRR value, VPP
+reference, trend job build ID and the LF testbed ID.
Jenkins Jobs
------------
@@ -202,34 +212,34 @@ data per test case. PT is designed as follows:
1. PT job triggers:
- a) Periodic e.g. daily.
+ a) Periodic e.g. twice a day.
b) On-demand gerrit triggered.
2. Measurements and data calculations per test case:
- a) Max Received Rate (MRR) - send packets at link rate over a trial
- period, count total received packets, divide by trial period.
+ a) Max Received Rate (MRR) - for each trial measurement,
+ send packets at link rate for trial duration,
+ count total received packets, divide by trial duration.
-3. Archive MRR per test case.
+3. Archive MRR values per test case.
4. Archive all counters collected at MRR.
Performance Analysis (PA)
`````````````````````````
-CSIT PA runs performance analysis including trendline calculation, trend
-compliance and anomaly detection using specified trend analysis metrics
-over the rolling window of last <N> sets of historical measurement data.
+CSIT PA runs performance analysis
+including anomaly detection as described above.
PA is defined as follows:
1. PA job triggers:
- a) By PT job at its completion.
+ a) By PT jobs at their completion.
b) On-demand gerrit triggered.
2. Download and parse archived historical data and the new data:
a) Download RF output.xml files from latest PT job and compressed
- archived data.
+ archived data from nexus.
b) Parse out the data filtering test cases listed in PA specification
(part of CSIT PAL specification file).
@@ -238,17 +248,21 @@ PA is defined as follows:
4. Evaluate new test data:
a) If the existing group is prolonged => Result = Pass,
- Reason = Normal. (to be updated base on the final Jenkins code).
- b) If a new group is detected with lower average => Result = Fail, Reason = Regression.
- c) If a new group is detected with higher average => Result = Pass, Reason = Progression.
+ Reason = Normal.
+ b) If a new group is detected with lower average =>
+ Result = Fail, Reason = Regression.
+ c) If a new group is detected with higher average =>
+ Result = Pass, Reason = Progression.
5. Generate and publish results
- a) Relay evaluation result to job result. (to be updated base on the
- final Jenkins code).
- b) Generate a new set of trend summary dashboard and graphs.
+ a) Relay evaluation result to job result.
+ b) Generate a new set of trend summary dashboard, list of failed
+ tests and graphs.
c) Publish trend dashboard and graphs in html format on
https://docs.fd.io/.
+ d) Generate an alerting email. This email is sent by Jenkins to
+ csit-report@lists.fd.io
Testbed HW configuration
------------------------