Channels
doorman
infrastructure
random
zercurity
community-feeds
fleet-dev
code-review
queryhub
apple-silicon
carving
tls
fim
goquery
zentral
aws
querycon
golang
zeek
file-carving
fuzzing
auditing-warroom
linen-dev
fleetosquery
plugins
jobs
arm-architecture
darkbytes
process-auditing
uptycs
android_tests
selfgroup
vendor-feeds
fleet
eclecticiq-polylogyx-extension
ebpf
website
core
general
macos
kolide
osctrl
extensions
foundation
sql
officehours
linux
windows
Title
a
arod
04/15/2021, 11:26 PMHey folks.
Is there a chart that talks resources to provide fleet server versus how many agents are reporting back?
n
Noah Talerman
04/16/2021, 3:09 PMHi @arod. Do you mind explaining a bit more what you mean by “talks resources to provide fleet server?” I believe the answer to your question is no, there is currently no such chart.
I’d like to learn more about what the ideal (most helpful) chart looks like for this resources use case.
a
arod
04/16/2021, 3:25 PMSure thing
Something like
Number of Osquery Agents | CPU Recommended | RAM Recommended
100. | 4vCPU. |. 8GB RAMJust trying to understand what fleet is expecting as far as food 🙂
As it gets higher, there's a lot of speculation of course. Which everyone should understand
z
zwass
04/16/2021, 4:45 PMThis is something that we would like to document but haven't done much of yet. If you give some details on your (potential) deployment we can give an estimate on what might be good.
a
arod
04/16/2021, 5:38 PM@zwass
No worries. Thanks. I can provide feedback based on the different environments we see once we start deploying the builds.
What are your thoughts on the following 3 use-cases
50 endpoints
100 endpoints
250 endpoints
Various OSes. Fleet only having a few scheduled packs per OS. Nothing crazy.
z
zwass
04/16/2021, 5:43 PMAt those scales you could have a very small instance 1vCPU/500MB RAM would do it for the Fleet server. You could probably even have the Redis and MySQL on the same machine if you want to optimize for cost and don't need failover, etc.
🙌 1