KPrasch
8dbaaaa8fd
Apply suggestions from code review
...
Co-authored-by: Derek Pierre <derek.pierre@gmail.com>
2024-01-23 13:27:35 +01:00
Manuel Montenegro
253a2067e3
Remove unused metrics
2024-01-23 13:27:35 +01:00
Manuel Montenegro
9512d7b286
Convert node discovery metrics from enum to gauge
...
Actually, the two possible states for learning_status metric are
`stopped` and `running`. So a enum type metric can be considered an
overkill that introduces confusion to exported metrics.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
933a28697d
Add metrics for root and child networks
...
Up to now, there is any distinction between root and child networks, so
the only metrics about Blockchain was related to root network.
This change splits the metrics into two networks.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
920df0e47d
Remove Operator and Event metrics
...
- Event metrics collector was not being used at all, so we can safely
remove it.
- Operator metrics collector is exporting the operator eth balance,
which is not relevant anymore.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
4cca43c0a6
Format code
2024-01-23 13:27:35 +01:00
Manuel Montenegro
e6c5293779
Move imports to the top of the file
...
Since commit de85e7fa
, prometheus is a required dependency for Ursula.
So conditional imports of prometheus are no longer necessary.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
13948b2625
Remove unused code and TODO
2024-01-23 13:27:35 +01:00
Manuel Montenegro
6a72807aec
Add Prometheus metrics endpoint to running logs
2024-01-23 13:27:35 +01:00
Manuel Montenegro
e51d637706
Add integration test for prometheus exporter
...
Test if start_prometheus_exporter has been properly called when running
Ursula.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
8c914ba30f
Remove redundant code in test_prometheus.py
2024-01-23 13:27:35 +01:00
Manuel Montenegro
2e6ed2f29c
Add CLI tests for Prometheus
2024-01-23 13:27:35 +01:00
Manuel Montenegro
7f82edd7ab
Make Prometheus exporter always run along Ursula
...
Prometheus metrics is going to be a necessary part for the Threshold
network. Through them, the network can know if the PRE nodes are meeting
the nodes requirement. This can be useful, for instance, to calculate
the staking rewards.
So, the export of these metrics must be mandatory and not an option that
the user can enable or disable using commands flags.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
78bbeb7dc1
Add mock for prometheus exporter start function
...
Since Prometheus exporter has become a non-optional part of Ursula, the
function start_prometheus_exporter is always called when running a new
Ursula.
Some tests, like test_ursula_run_ip_checkup runs several Ursulas, so
this function is called multiple times.
Only one instance of Prometheus exporter must be run at same time, so
this make the tests to fail because some Prometheus errors are raised.
Co-authored-by: LunarBytes <kieran@nucypher.com>
2024-01-23 13:27:35 +01:00
Kieran Prasch
d13c398278
newsfrag for PR #3310
2024-01-23 13:27:35 +01:00
Kieran Prasch
e81289ab18
Relock dependencies with prometheus-client as a primary dependency.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
7f70c9a086
Disable default collector metrics
...
By default, prometheus exports 'process', 'gc' and 'platform' collector
metrics. This information is not useful in our case and introduces some
noise to the metrics, so they have been disabled.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
11f973a97a
Rename host Prometheus metrics
...
- host_info has been renamed to client_info
- node_discovery has been renamed to node_discovery_status
- New field for client_info: `app: TACo`
2024-01-23 13:27:35 +01:00
Manuel Montenegro
40628ae991
Remove metrics prefix
2024-01-23 13:27:35 +01:00
Manuel Montenegro
ac1893cd50
Remove optional imports of Prometheus
...
Prometheus is now a dependency that is always installed along with
nucypher. For this reason, it doesn't make sense to check if Prometheus
is installed before importing it.
2024-01-23 13:27:35 +01:00
Manuel Montenegro
97ab6bc23b
updates setup.py
2024-01-23 13:27:35 +01:00
KPrasch
2ae6ce85b0
Merge pull request #3393 from cygnusv/nojoin
...
Optimize use of decryption request WorkerPool
2024-01-19 15:01:36 +01:00
David Núñez
5212834c38
newsfragment for 3393
2024-01-19 12:54:25 +01:00
David Núñez
55e653cf75
Tweak WorkerPool parameters
2024-01-19 12:13:15 +01:00
derekpierre
a508765671
Add newsfragment for #3333 .
2024-01-19 11:04:31 +01:00
derekpierre
4100754976
Raise exception for sampling when there aren't enough nodes to sample.
2024-01-19 11:04:31 +01:00
derekpierre
76b9d4af28
Add unit tests for testing MergedReservoir and PrefetchStrategy.
2024-01-19 11:04:31 +01:00
derekpierre
bad425f17f
Raise exception when the reservoir of values to draw from is less than required successes.
2024-01-19 11:04:31 +01:00
Piotr Roslaniec
b0f9fe1897
fix darker linting
2024-01-18 12:41:45 +01:00
Piotr Roslaniec
d001eab84b
add a newsfragment
2024-01-18 12:41:45 +01:00
Piotr Roslaniec
04b3202877
set MAX_UPLOAD_CONTENT_LENGTH to 250kb
2024-01-18 12:41:45 +01:00
David Núñez
deb3391131
Increment DEFAULT_DECRYPTION_TIMEOUT to 30 sec
2024-01-17 16:32:52 +01:00
David Núñez
773fa467a6
Reraise exceptions inside TDec worker tasks as ThresholdDecryptionRequestFailed
2024-01-17 16:32:00 +01:00
David Núñez
a7af04beb8
Randomize sequence of addresses for decryption requests
2024-01-17 16:26:55 +01:00
David Núñez
d0c25cf8d4
Include batch_size as a parameter of ThresholdDecryptionRequestFactory
2024-01-17 15:35:12 +01:00
derekpierre
a8c68109ab
Add auto-use fixture to limit maximum post aggregation delay.
2024-01-17 09:31:32 -05:00
derekpierre
5163aed044
Use ritual id 5 for lynx examples now that ritual 0 is expired.
2024-01-17 09:31:32 -05:00
derekpierre
be8820786f
Implement a random stagger delay when nodes submit aggregated transcripts.
2024-01-17 09:31:32 -05:00
derekpierre
8a341dd87c
Use an increased modifier for posting aggregation in case node is the last node and needs more gas for additoinal logic of being the last node to submit aggregated transcript.
...
Basically a race condition for who will be last.
2024-01-17 09:31:32 -05:00
derekpierre
0d71361f88
Have the EventScanner take a break after crashing and then restarting.
2024-01-17 09:31:32 -05:00
derekpierre
2426bacef1
Add bugfix newsfragment for #3390 .
2024-01-17 09:31:32 -05:00
derekpierre
c5bf021c24
Update tests to have ferveo public keys obtained from CoordinatorAgent/Coordinator contract.
...
Co-authored-by: Kieran Prasch <kieranprasch@gmail.com>
2024-01-17 09:31:32 -05:00
derekpierre
172fefe74f
Obtain node ferveo public key from Coordinator contract instead of learning.
...
Co-authored-by: Kieran Prasch <kieranprasch@gmail.com>
2024-01-17 09:31:32 -05:00
David Núñez
7e8caefabb
Fix parameter typing
2024-01-17 12:56:16 +01:00
David Núñez
dedf747fe2
Fix incorrect exception message
2024-01-17 12:51:44 +01:00
David Núñez
70d7fe1319
Most importantly, no need to wait for all threads to finalize
...
This was making the whole process as slow as the slowest decryption request
2024-01-17 12:51:12 +01:00
David Núñez
e19dd65b3c
Wait enough time before firing a new batch of decryption requests
2024-01-17 12:50:26 +01:00
David Núñez
30a39c59d8
Batch size should start with a bit over the threshold, just in case
2024-01-17 12:48:49 +01:00
David Núñez
43df0adf7b
Fix parameter name
2024-01-17 12:47:02 +01:00
derekpierre
c0a59d48d7
Bump version: 7.0.4 → 7.1.0
2024-01-10 11:56:45 -05:00