I'm looking at the possibility of building our engineering lab into AWS
where we will be charged 24/7, but my individual engineers don't work 24/7
so time-based scaling. Anyone else working with this use case? It mostly
works. I got the idea from moua in #scalr IRC.
Inside the Farm Role:
- Set Minimum instance = 0 and and Maximum > 0
- Add a scaling rules based on "DateAndTime"
- Schedule some rules e.g.
- 1 instance(s), between 7:45 AM and 5:00 PM on Mon, Tue, Wed, Thu,
Fri
- 0 instance(s), between 5:15 PM and 11:45 PM on Mon, Tue, Wed, Thu,
Fri
- 0 instance(s), between 12:00 AM and 7:30 AM on Mon, Tue, Wed, Thu,
Fri
- weekend rule... etc.
- Set "Termination Preferences"
- Scaling Behavior = Resume / Suspend
- Consider Suspended Server = Terminated
- Skip waiting the full billing period
Done. Now Lab servers suspend after work hours and spin up before the
workday. Now I have a Farm that's 1/4 of the cost.
*PROBLEMS*
Really, I have three right now.
- *MYSQL ROLE FAILS* - I built this in the "3-Tiered App" Farm and it
works for NGINX and Apache, but not for MYSQL. It just doesn't shutdown.
- *REPLICATION* - It's a little bit labor intensive so setup and its
possible I'll have dozens of Farms. Any ideas on how I could automate the
creation of Farms, or Scaling rules inside farms?
- *BUDGET PROJECTION* - Any screen that makes a guess about my budget
does with 100% utilization in mind. I don't really care too much about
this, but thought I'd mention it.
Thanks... and I really enjoy working with Scalr.
--
You received this message because you are subscribed to the Google Groups
"scalr-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
