6ae24b1ff9
use cache's expiration and race-free get_with
...
when this was a dashmap, we needed our own expiration and parallel requests would do the same query.
with moka, we can use their expiration code and get_with
2022-09-20 01:33:39 +00:00
90fed885bc
add ttl to user_cache
2022-09-19 23:09:31 +00:00
8785cb6d2d
no idle timeout or max lifetime
2022-09-19 22:41:53 +00:00
b2adc73b50
upgrade tokio-stream
2022-09-19 22:25:45 +00:00
3e3a9fcf64
larger max_capacity now that there is a weigher
2022-09-19 22:24:56 +00:00
a7761a0430
add a weigher function to sized caches
2022-09-19 22:17:35 +00:00
b6275aff1e
no need for an atomic bool
2022-09-19 22:17:24 +00:00
28fa424c2a
start weighers
2022-09-17 02:30:06 +00:00
05af0551c9
use ahash
...
this might actually be slower. do real benchmarks
2022-09-17 02:17:20 +00:00
3d4bfbfde0
use ahash. allow no redis
2022-09-17 01:19:11 +00:00
12b6d01434
work in progress
2022-09-15 17:57:24 +00:00
7a1433e3c9
no longer need to use total difficulty on ETH 2.0
2022-09-14 19:39:08 +00:00
31227d8c0c
drop in flight. throughput is enough
2022-09-14 07:08:48 +00:00
731a255d2f
add old block to log and more todos
2022-09-14 06:57:24 +00:00
ea8f6091f9
improve (but disable) rate limits
2022-09-14 06:41:34 +00:00
ed505cdd23
fix check for http providers
...
even the websocket providers have an http_client set. maybe drop that?
2022-09-14 06:38:53 +00:00
1730b8c7a5
switch to deadpool
2022-09-14 06:18:13 +00:00
789672be43
try fixing warning about missing blocks
2022-09-14 05:26:46 +00:00
a5ccec76c7
silence some warnings and make private rpcs an option
2022-09-14 04:27:18 +00:00
1d4d5844d0
reconnect -> connect
2022-09-14 03:32:38 +00:00
612f5a48c1
minor polish to error
2022-09-14 03:01:58 +00:00
5a16b9eed8
fix reconnect for http clients
2022-09-14 02:11:48 +00:00
7eaf6f3540
retrying reconnect
2022-09-14 01:43:09 +00:00
73a2fcaf72
add FAQ
2022-09-14 01:42:48 +00:00
519c312e8d
reconnect needs to clear more state
2022-09-12 23:00:10 +00:00
dc30bdcef3
cargo upgrade
2022-09-12 14:36:26 +00:00
05e94ff81c
send UNAUTHORIZED to unknown keys
2022-09-12 14:33:55 +00:00
7ff319e9b0
reconnect sooner
2022-09-12 14:33:19 +00:00
88661fa865
return UnknownKey instead of error
2022-09-10 03:46:02 +00:00
d98886db14
simple status json and prometheus
2022-09-10 02:59:07 +00:00
501f2b0b24
improve logging
2022-09-10 00:58:33 +00:00
879c6e49f2
dry errors so that rate limits dont log so much
2022-09-10 00:12:14 +00:00
c8da98d12e
even more metrics
2022-09-09 03:53:16 +00:00
cae034afb3
better metrics and spawn
2022-09-08 21:01:36 +00:00
cfca16319b
lower uncle log level
2022-09-07 20:24:45 +00:00
2c41cad452
more counts to the status page
2022-09-07 04:11:47 +00:00
c9b2c0c0d2
dry cache code
2022-09-07 03:54:16 +00:00
5eef5173a1
disable redis rate limits
2022-09-06 22:55:17 +00:00
e4d25b207d
better default connection counts
2022-09-06 20:50:37 +00:00
c89295ef7e
we should use retry_at
2022-09-06 20:33:25 +00:00
c34e8ef383
improve rate limiting and request counters
2022-09-06 20:13:07 +00:00
98265424bb
set response id in another spot. this is too fragile
2022-09-06 17:56:49 +00:00
041dd5f317
more logs and beginning of a health check
2022-09-06 16:49:07 +00:00
83a7b03dea
improve no-op subscription
...
can still be improved more
2022-09-06 13:14:15 +00:00
bb96757452
None instead of Block::default() more places
2022-09-06 12:29:37 +00:00
e8977d203b
use None more
2022-09-06 03:26:23 +00:00
4502344027
add head_block_id to connection serializer
2022-09-05 16:39:46 +00:00
8080794a8b
skip dupes earlier
2022-09-05 16:29:21 +00:00
068c05cf4f
improve fork logic again
2022-09-05 16:25:21 +00:00
2092e74fd3
fix variable names and log to match
2022-09-05 15:49:53 +00:00
Bryan Stitt