hubble
A public mirror for the whole atmosphere hubble.microcosm.blue
Star
Fork
Watch

Commits on backup

phil
bump eat-rocks (faster incremental restore)
committed
phil
fix: don't create the checkpoint dir, rocks does
committed
phil
configurable initial backup delay
committed
phil
initial eat-rocks push
committed
phil
bump deps
committed
phil
bump a few deps
committed
phil
bump versions for release
committed
phil
oops skip slots
committed
phil
Merge commit 'edcab0b0ddf79ab5e2e93be0cac8f346307f1afb'
committed
phil
out-of-scope reevaluation: scan on startup
committed
phil
debug: where are we losing the commits
committed
phil
store `prev` from the commit object
committed
phil
configurable archive export compression
committed
phil
update todos
committed
phil
turn off open admin mod writes for now lol
committed
phil
Merge branch 'bridgy-auth-getrepo-offload'
committed
phil
wire in the deep crawl
committed
phil
deep crawl
committed
phil
show did instead of index on error bridgy-auth-getrepo-offload
committed
phil
handle response encoding (compression) in hubble
committed
phil
conditionally install tokio-console subscriber
committed
phil
buffer more output when writing archives
committed
phil
track in-flight over response streaming
committed
phil
getRepo bench
committed
phil
split limit waits from request times
committed
phil
opt-in auth token forward-across-redirect thing
committed
phil
tolerate missing tooBig on #commit
committed
phil
configs for server binds + upstream limits
committed
phil
add hubble request metrics
committed
phil
fix test name
committed
phil
send an auth token with getRepo (optionally)
committed
phil
resync queue guages
committed
phil
get more timing info for resyncs
committed
phil
blehhhhhh
committed
phil
tokio-console
committed
phil
Revert "please let this not work"
committed
phil
please let this not work
committed
phil
i guess it's this poll timeout...
committed
phil
one more micro-optimization
committed
phil
micro-optmize mushrooms
committed
phil
maybe 16?
committed
phil
moar mushroom requests at a time
committed
phil
batch-prefetch queued resyncs
committed
phil
bypass relay for getRepo on first-seen
committed
phil
ok chill on the relay + allow more open files
committed
phil
bump relay concurrency
committed
phil
evict more actors at a time
committed
phil
tests for record did interning
committed
phil
use interned dids for records
committed
phil
big machine bit faster
committed
phil
one memtable per flush
committed
phil
bleh. plz faster.
committed
phil
go faster
committed
phil
metrics for resynced bytes by ingest/batch-put
committed
phil
shutdown: don't wait on compactions, log
committed
phil
downgrade file-remove fail notice (it's normal)
committed
phil
oop: create sst staging dir
committed
phil
sure, compression regression check
committed
phil
straggler
committed
phil
oop, match the sst ingest config to bottommost
committed