def build_span_tree(span_tree, spans, parent_span_id):
for span_id, child in sorted(span_tree.items(), key=lambda item: item[0]):
span = copy.deepcopy(reference_span)
# non-leaf node span
span["parent_span_id"] = parent_span_id.ljust(16, "0")
span["span_id"] = span_id.ljust(16, "0")
(start_delta, span_length) = time_offsets.get(span_id, (timedelta(), timedelta()))
span_start_time = start_datetime + start_delta
span["start_timestamp"] = timestamp_format(span_start_time)
span["timestamp"] = timestamp_format(span_start_time + span_length)
spans.append(span)
if isinstance(child, dict):
spans = build_span_tree(child, spans, span_id)
elif isinstance(child, six.string_types):
parent_span_id = span_id
span_id = child
span = copy.deepcopy(reference_span)
# leaf node span
span["parent_span_id"] = parent_span_id.ljust(16, "0")
span["span_id"] = span_id.ljust(16, "0")
(start_delta, span_length) = time_offsets.get(span_id, (timedelta(), timedelta()))
span_start_time = start_datetime + start_delta
span["start_timestamp"] = timestamp_format(span_start_time)
span["timestamp"] = timestamp_format(span_start_time + span_length)
spans.append(span)
return spans
def regress_spans(original_spans):
spans = []
last_index = len(original_spans) - 1
for index, reference_span in enumerate(original_spans):
end_datetime = datetime.utcfromtimestamp(
reference_span["timestamp"]).replace(tzinfo=pytz.utc)
span = copy.deepcopy(reference_span)
if index % 2 == 0:
# for every even indexed span, increase its duration.
# this implies the span was slower.
regression_time = timedelta(milliseconds=10 + index)
span["timestamp"] = timestamp_format(end_datetime +
regression_time)
else:
# for every odd indexed span, decrease its duration.
# this implies the span was faster
regression_time = timedelta(milliseconds=5 + index)
span["timestamp"] = timestamp_format(end_datetime -
regression_time)
if index == last_index:
# change the op name of the last span.
# the last span would be removed from the baseline transaction;
# and the last span of the baseline transaction would be missing from
# the regression transaction
span["op"] = "resource"
spans.append(span)
return spans
def test_transaction(self):
event_data = {
"event_id":
"d2132d31b39445f1938d7e21b6bf0ec4",
"type":
"transaction",
"transaction":
"/organizations/:orgId/performance/:eventSlug/",
"start_timestamp":
iso_format(before_now(minutes=1, milliseconds=500)),
"timestamp":
iso_format(before_now(minutes=1)),
"contexts": {
"trace": {
"trace_id": "ff62a8b040f340bda5d830223def1d81",
"span_id": "8f5a2b8768cafb4e",
"type": "trace",
}
},
"spans": [
{
"description":
"<OrganizationContext>",
"op":
"react.mount",
"parent_span_id":
"8f5a2b8768cafb4e",
"span_id":
"bd429c44b67a3eb4",
"start_timestamp":
timestamp_format(before_now(minutes=1, milliseconds=250)),
"timestamp":
timestamp_format(before_now(minutes=1)),
"trace_id":
"ff62a8b040f340bda5d830223def1d81",
},
{
"description":
"browser span",
"op":
"browser",
"parent_span_id":
"bd429c44b67a3eb4",
"span_id":
"a99fd04e79e17631",
"start_timestamp":
timestamp_format(before_now(minutes=1, milliseconds=200)),
"timestamp":
timestamp_format(before_now(minutes=1)),
"trace_id":
"ff62a8b040f340bda5d830223def1d81",
},
{
"description":
"resource span",
"op":
"resource",
"parent_span_id":
"bd429c44b67a3eb4",
"span_id":
"a71a5e67db5ce938",
"start_timestamp":
timestamp_format(before_now(minutes=1, milliseconds=200)),
"timestamp":
timestamp_format(before_now(minutes=1)),
"trace_id":
"ff62a8b040f340bda5d830223def1d81",
},
{
"description":
"http span",
"op":
"http",
"parent_span_id":
"a99fd04e79e17631",
"span_id":
"abe79ad9292b90a9",
"start_timestamp":
timestamp_format(before_now(minutes=1, milliseconds=200)),
"timestamp":
timestamp_format(before_now(minutes=1)),
"trace_id":
"ff62a8b040f340bda5d830223def1d81",
},
{
"description":
"db span",
"op":
"db",
"parent_span_id":
"abe79ad9292b90a9",
"span_id":
"9c045ea336297177",
"start_timestamp":
timestamp_format(before_now(minutes=1, milliseconds=200)),
"timestamp":
timestamp_format(before_now(minutes=1)),
"trace_id":
"ff62a8b040f340bda5d830223def1d81",
},
],
}
with Feature({
"organizations:performance-ops-breakdown": True,
"organizations:performance-suspect-spans-ingestion": True,
}):
event = self.post_and_retrieve_event(event_data)
raw_event = event.get_raw_data()
exclusive_times = [
pytest.approx(50),
pytest.approx(0),
pytest.approx(200),
pytest.approx(0),
pytest.approx(200),
]
assert raw_event["spans"] == [
dict(span,
exclusive_time=exclusive_time,
hash=hash_values([span["description"]])) for span,
exclusive_time in zip(event_data["spans"], exclusive_times)
]
assert raw_event["breakdowns"] == {
"span_ops": {
"ops.browser": {
"value": pytest.approx(200)
},
"ops.resource": {
"value": pytest.approx(200)
},
"ops.http": {
"value": pytest.approx(200)
},
"ops.db": {
"value": pytest.approx(200)
},
"total.time": {
"value": pytest.approx(1050)
},
}
}
