fix(ci): Ensure integration-tests allow for at least 30 seconds to establish a connection (#3676)

So the cause of the flaky tests is that they aren't waiting long enough
for a connection to be established. Both the test in #3666 and the
`iperf` tests have a timeout of 10 seconds.

Connections _should_ be established **very quickly** in CI. However, I
have a few guesses as to why they might not be, essentially causing us
to have to wait for a timeout to re-initiate a connection request:

- Packets arrive out of order or too quickly for the WireGuard state
machine to establish a handshake.
- Too many ICE candidates gathered (the gateway has 3 interfaces)


This PR:

- Refactors the iperf tests to be a little easier to maintain
- Ensures `integration-tests` run for at least 30 seconds before timing
out


In any case, we can debug / optimize this further after snownet is
merged, which might just solve the problem completely.
This commit is contained in:
Jamil
2024-02-19 12:50:58 -08:00
committed by GitHub
parent 4dc8cdf908
commit eebd7fc7f1
4 changed files with 220 additions and 147 deletions

View File

@@ -243,13 +243,11 @@ jobs:
include:
- test_name: direct-perf
setup: echo 'Noop'
execute: echo 'Noop'
- test_name: relayed-perf
setup: |
# Disallow traffic between gateway and client container
sudo iptables -I FORWARD 1 -s 172.28.0.100 -d 172.28.0.105 -j DROP
sudo iptables -I FORWARD 1 -s 172.28.0.105 -d 172.28.0.100 -j DROP
execute: echo 'Noop'
steps:
- uses: actions/checkout@v4
@@ -277,41 +275,15 @@ jobs:
- name: 'Setup test: ${{ matrix.test_name }}'
run: ${{ matrix.setup }}
- name: 'Execute test: ${{ matrix.test_name }}'
run: ${{ matrix.execute }}
- name: 'Performance test: ${{ matrix.test_name }}'
id: perfomance-test
id: performance-test
timeout-minutes: 5
run: |
set -xe
mkdir -p /tmp/iperf3results
docker compose exec --env RUST_LOG=info -it client /bin/sh -c 'iperf3 -O 1 -R -c 172.20.0.110 --json' >> /tmp/iperf3results/tcp_server2client.json
cat /tmp/iperf3results/tcp_server2client.json | jq -r '"tcp_server2client_sum_received_bits_per_second=" + (.end.sum_received.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/tcp_server2client.json | jq -r '"tcp_server2client_sum_sent_bits_per_second=" + (.end.sum_sent.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/tcp_server2client.json | jq -r '"tcp_server2client_sum_sent_retransmits=" + (.end.sum_sent.retransmits|tostring)' >> "$GITHUB_OUTPUT"
docker compose exec --env RUST_LOG=info -it client /bin/sh -c 'iperf3 -O 1 -c 172.20.0.110 --json' >> /tmp/iperf3results/tcp_client2server.json
cat /tmp/iperf3results/tcp_client2server.json | jq -r '"tcp_client2server_sum_received_bits_per_second=" + (.end.sum_received.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/tcp_client2server.json | jq -r '"tcp_client2server_sum_sent_bits_per_second=" + (.end.sum_sent.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/tcp_client2server.json | jq -r '"tcp_client2server_sum_sent_retransmits=" + (.end.sum_sent.retransmits|tostring)' >> "$GITHUB_OUTPUT"
# Note: birtate is reduced to be 250M but what we actually want to test for is 1G once we flesh out some bugs
docker compose exec --env RUST_LOG=info -it client /bin/sh -c 'iperf3 -O 1 -u -b 250M -R -c 172.20.0.110 --json' >> /tmp/iperf3results/udp_server2client.json
cat /tmp/iperf3results/udp_server2client.json | jq -r '"udp_server2client_sum_bits_per_second=" + (.end.sum.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/udp_server2client.json | jq -r '"udp_server2client_sum_jitter_ms=" + (.end.sum.jitter_ms|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/udp_server2client.json | jq -r '"udp_server2client_sum_lost_percent=" + (.end.sum.lost_percent|tostring)' >> "$GITHUB_OUTPUT"
# Note: birtate is reduced to be 250M but what we actually want to test for is 1G once we flesh out some bugs
docker compose exec --env RUST_LOG=info -it client /bin/sh -c 'iperf3 -O 1 -u -b 250M -c 172.20.0.110 --json' >> /tmp/iperf3results/udp_client2server.json
cat /tmp/iperf3results/udp_client2server.json | jq -r '"udp_client2server_sum_bits_per_second=" + (.end.sum.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/udp_client2server.json | jq -r '"udp_client2server_sum_jitter_ms=" + (.end.sum.jitter_ms|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results/udp_client2server.json | jq -r '"udp_client2server_sum_lost_percent=" + (.end.sum.lost_percent|tostring)' >> "$GITHUB_OUTPUT"
run: ./scripts/tests/perf/test.sh
- name: 'Save performance test results: ${{ matrix.test_name }}'
uses: actions/upload-artifact@v4
with:
name: '${{ matrix.test_name }}-iperf3results'
path: /tmp/iperf3results
path: ./iperf3results
- name: 'Download main branch performance test results: ${{ matrix.test_name }}'
id: download-artifact
if: ${{ github.event_name == 'pull_request' }}
@@ -323,7 +295,7 @@ jobs:
REPO="${{ github.repository }}"
WORKFLOW="cd.yml"
ARTIFACT_NAME="${{ matrix.test_name }}-iperf3results"
DESTINATION="/tmp/iperf3results-main"
DESTINATION="./iperf3results-main"
ARTIFACTS_URL=$(
gh api \
@@ -347,125 +319,15 @@ jobs:
set -x
unzip "${DESTINATION}.zip" -d "${DESTINATION}"
rm "${DESTINATION}.zip"
- name: "Generate main branch metrics"
id: main-perfomance-test
if: ${{ github.event_name == 'pull_request' }}
run: |
cat /tmp/iperf3results-main/tcp_server2client.json | jq -r '"tcp_server2client_sum_received_bits_per_second=" + (.end.sum_received.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/tcp_server2client.json | jq -r '"tcp_server2client_sum_sent_bits_per_second=" + (.end.sum_sent.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/tcp_server2client.json | jq -r '"tcp_server2client_sum_sent_retransmits=" + (.end.sum_sent.retransmits|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/tcp_client2server.json | jq -r '"tcp_client2server_sum_received_bits_per_second=" + (.end.sum_received.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/tcp_client2server.json | jq -r '"tcp_client2server_sum_sent_bits_per_second=" + (.end.sum_sent.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/tcp_client2server.json | jq -r '"tcp_client2server_sum_sent_retransmits=" + (.end.sum_sent.retransmits|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_server2client.json | jq -r '"udp_server2client_sum_bits_per_second=" + (.end.sum.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_server2client.json | jq -r '"udp_server2client_sum_jitter_ms=" + (.end.sum.jitter_ms|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_server2client.json | jq -r '"udp_server2client_sum_lost_percent=" + (.end.sum.lost_percent|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_client2server.json | jq -r '"udp_client2server_sum_bits_per_second=" + (.end.sum.bits_per_second|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_client2server.json | jq -r '"udp_client2server_sum_jitter_ms=" + (.end.sum.jitter_ms|tostring)' >> "$GITHUB_OUTPUT"
cat /tmp/iperf3results-main/udp_client2server.json | jq -r '"udp_client2server_sum_lost_percent=" + (.end.sum.lost_percent|tostring)' >> "$GITHUB_OUTPUT"
- name: Update PR
- name: Update PR with results
uses: actions/github-script@v7
id: perf-comment
if: ${{ github.event_name == 'pull_request' }}
with:
github-token: ${{ secrets.GITHUB_TOKEN }}
script: |
// 1. Retrieve existing bot comments for the PR
const { data: comments } = await github.rest.issues.listComments({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.issue.number,
});
const botComment = comments.find(comment => {
return comment.user.type === 'Bot' && comment.body.includes('${{ matrix.test_name }}')
});
function humanFileSize(bytes, dp=1) {
const thresh = 1000;
if (Math.abs(bytes) < thresh) {
return bytes + ' B';
}
const units = ['KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'];
let u = -1;
const r = 10**dp;
do {
bytes /= thresh;
++u;
} while (Math.round(Math.abs(bytes) * r) / r >= thresh && u < units.length - 1);
return bytes.toFixed(dp) + ' ' + units[u];
}
function getDiffPercents(main, current)
{
let diff = -1 * (100 - current / (main / 100));
if (diff > 0) {
return "+" + diff.toFixed(0) + "%";
} else if (diff == 0) {
return "0%";
} else {
return diff.toFixed(0) + "%";
}
}
let tcp_server2client_sum_received_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.tcp_server2client_sum_received_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_server2client_sum_received_bits_per_second }}, ${{ steps.perfomance-test.outputs.tcp_server2client_sum_received_bits_per_second }}) + ')';
let tcp_server2client_sum_sent_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.tcp_server2client_sum_sent_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_server2client_sum_sent_bits_per_second }}, ${{ steps.perfomance-test.outputs.tcp_server2client_sum_sent_bits_per_second }}) + ')';
let tcp_server2client_sum_sent_retransmits = ${{ steps.perfomance-test.outputs.tcp_server2client_sum_sent_retransmits }} + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_server2client_sum_sent_retransmits }}, ${{ steps.perfomance-test.outputs.tcp_server2client_sum_sent_retransmits }}) + ')';
let tcp_client2server_sum_received_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.tcp_client2server_sum_received_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_client2server_sum_received_bits_per_second }}, ${{ steps.perfomance-test.outputs.tcp_client2server_sum_received_bits_per_second }}) + ')';
let tcp_client2server_sum_sent_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.tcp_client2server_sum_sent_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_client2server_sum_sent_bits_per_second }}, ${{ steps.perfomance-test.outputs.tcp_client2server_sum_sent_bits_per_second }}) + ')';
let tcp_client2server_sum_sent_retransmits = ${{ steps.perfomance-test.outputs.tcp_client2server_sum_sent_retransmits }} + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.tcp_client2server_sum_sent_retransmits }}, ${{ steps.perfomance-test.outputs.tcp_client2server_sum_sent_retransmits }}) + ')';
let udp_server2client_sum_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.udp_server2client_sum_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_server2client_sum_bits_per_second }}, ${{ steps.perfomance-test.outputs.udp_server2client_sum_bits_per_second }}) + ')';
let udp_server2client_sum_jitter_ms = (${{ steps.perfomance-test.outputs.udp_server2client_sum_jitter_ms }}).toFixed(2) + "ms (" + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_server2client_sum_jitter_ms }}, ${{ steps.perfomance-test.outputs.udp_server2client_sum_jitter_ms }}) + ')';
let udp_server2client_sum_lost_percent = (${{ steps.perfomance-test.outputs.udp_server2client_sum_lost_percent }}).toFixed(2) + "% (" + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_server2client_sum_lost_percent }}, ${{ steps.perfomance-test.outputs.udp_server2client_sum_lost_percent }}) + ')';
let udp_client2server_sum_bits_per_second = humanFileSize(${{ steps.perfomance-test.outputs.udp_client2server_sum_bits_per_second }}) + ' (' + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_client2server_sum_bits_per_second }}, ${{ steps.perfomance-test.outputs.udp_client2server_sum_bits_per_second }}) + ')';
let udp_client2server_sum_jitter_ms = (${{ steps.perfomance-test.outputs.udp_client2server_sum_jitter_ms }}).toFixed(2) + "ms (" + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_client2server_sum_jitter_ms }}, ${{ steps.perfomance-test.outputs.udp_client2server_sum_jitter_ms }}) + ')';
let udp_client2server_sum_lost_percent = (${{ steps.perfomance-test.outputs.udp_client2server_sum_lost_percent }}).toFixed(2) + "% (" + getDiffPercents(${{ steps.main-perfomance-test.outputs.udp_client2server_sum_lost_percent }}, ${{ steps.perfomance-test.outputs.udp_client2server_sum_lost_percent }}) + ')';
const output = `## Performance Test Results: ${{ matrix.test_name }}
### TCP
| Direction | Received/s | Sent/s | Retransmits |
|------------------|--------------------------------------------------------|----------------------------------------------------|------------------------------------------------|
| Client to Server | ` + tcp_client2server_sum_received_bits_per_second + ` | ` + tcp_client2server_sum_sent_bits_per_second + ` | ` + tcp_client2server_sum_sent_retransmits + ` |
| Server to Client | ` + tcp_server2client_sum_received_bits_per_second + ` | ` + tcp_server2client_sum_sent_bits_per_second + ` | ` + tcp_server2client_sum_sent_retransmits + ` |
### UDP
| Direction | Total/s | Jitter | Lost |
|------------------|-----------------------------------------------|-----------------------------------------|--------------------------------------------|
| Client to Server | ` + udp_client2server_sum_bits_per_second + ` | ` + udp_client2server_sum_jitter_ms + ` | ` + udp_server2client_sum_lost_percent + ` |
| Server to Client | ` + udp_server2client_sum_bits_per_second + ` | ` + udp_server2client_sum_jitter_ms + ` | ` + udp_client2server_sum_lost_percent + ` |
`;
// 3. Update previous comment or create new one
if (botComment) {
github.rest.issues.updateComment({
owner: context.repo.owner,
repo: context.repo.repo,
comment_id: botComment.id,
body: output
});
} else {
github.rest.issues.createComment({
issue_number: context.issue.number,
owner: context.repo.owner,
repo: context.repo.repo,
body: output
});
}
const { script } = require('./scripts/tests/perf/results.js');
script(github, context, '${{ matrix.test_name }}');
- name: Show Client logs
if: "!cancelled()"
run: docker compose logs client