|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. We're about to implement on-call for our team to handle the fact that, lacking on call, one engineer feels like he's handling the entire thing. Here's how we're setting it up: Is on-call mandatory for you? yes How often and for how long are you on-call? 1 week every 2 months. 2 folks per rotation, 16 total on call engineers (everyone on the team aside from a couple of Jr's that we're bringing along more slowly). EM's have their own escalation rotation What kind of duties are you expected to do for on-call? Watch the issues slack channel and triage. IF a notification comes through, during your working hours, triage that. How often does something happen while you're on-call?(on or off hours) 2-3 times /week Do you get paid anything additional for on-call? no The goal is to make it a light enough thing with the double rotation that it's not actually taking anyone away from work. I've had much different setups elsewhere.
|
#
?
Oct 31, 2023 19:36
|
|
|
|
| # ? May 25, 2024 23:41 |
|
|
On call is pretty tolerable if you're only dealing with P0 stuff and it breaks once a week. It's gonna ruin your whole day but it's not going to break your delivery times too much At places where on call is also daytime Also need to budget time for post mortem document prep and meetings which I count at about an hour per, on top of the actual P0
|
|
#
?
Oct 31, 2023 20:08
|
|
|
it is usefull to distinguish between oncall and onduty. what guinness called "oncall"...isn't. that's on duty. handling tickets and other interrupting tasks during business hours so the rest of the team can do project work. oncall is when it is not work-time, and yet you still have to handle things when something is down and they ... call you. for onduty, you are on interrupts/tickets/operational...duty. when your duty shift is over, you're done.
|
|
#
?
Oct 31, 2023 20:31
|
|
|
and with that distinction, im on operational duty for a week out of every six or eight. it's all during business hours, and ops work is expected to be p0. there is generally no expectation that i'll do anything for project work during that time. it's actually the case that half of my project work is aligning other teams and stuff, which means going to meetings. since i still need to take meetings when i'm on interrupts duty, some of my project work still gets done. but when i was doing more in-an-ide-writing-code type stuff, i would do no project work during my operational shift. leftover time during an ops shift should be spent reducing operational burden by automating processes, writing tooling/documentation to be more self-service, etc. ofc if the _only_ time ops stuff is automated is during ops rotation, you're in for a bad time. sometimes project time needs to be dedicated to reducing operational load, too. but eng time during an ops shift should almost always be dedicated to that imo
|
|
#
?
Oct 31, 2023 21:08
|
|
|
Onduty and oncall are the same person on my team, so that’s a sprint lost there.
|
|
#
?
Oct 31, 2023 23:30
|
|
|
Hadlock posted:DevOps jobs Fuckin A, dude. I’m doing what I can to keep up but nobody knows to leave well enough alone and chill out with adding unnecessary complexity. I’ve done plenty of excellent work in my career up to this point and I keep being expected to do more and more.
|
|
#
?
Nov 1, 2023 01:26
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. - Yes - Every 6 weeks, I have a one week rotation - Point of contact for all support, respond to alerts/pages when poo poo hits the fan - Alerts/pages are really infrequent these days. There's usually an incident bridge once every few weeks and we're usually just brought in as people potentially involved. - No We're a high priority team so really bad things can happen, but we've worked really hard in the last year to limit false alerts, fix bugs that drive incidents, etc. The best weeks there's essentially zero extra work. The worst weeks it's probably like 50-75% of your time, but those are pretty rare. When we used to point out our support shifts I seem to recall it ended up being like 20-25% of a dev's time that we estimated. luchadornado fucked around with this message at 02:00 on Nov 1, 2023 |
|
#
?
Nov 1, 2023 01:56
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. Weighing in for my current team but I've done much worse previously: Oncall is mandatory and a big part of our service offering (we are essentially DevOps/SRE) We have a primary and secondary role and split equally among around 8 people so approximately 1/4 of the time. Our hours are from around 10-6 depending on the time of year since we have teams in other countries. Primary handles incident response. Secondary handles lower severity stuff and customer reach outs etc, and change management. We don't do off hours but our oncall is generally considered fully utilized, especially secondary. Edit: we do weekends but just primary right now and it's generally pretty dead. No additional payment that I'm aware of but we are a faang so everyone is making FAANG money? Eh. Falcon2001 fucked around with this message at 02:39 on Nov 1, 2023 |
|
#
?
Nov 1, 2023 02:37
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. 1) Current job, no - but we have nothing in production so that is why. 2) In the past it has been an even rotation among all the engineers on the team/in the org depending on size. 3) Outside of work hours the expectation has generally been to respond to critical incidents. During work hours there has sometimes been an expectation that you are spending your time on the rotation improving observability/fixing minor bugs or working on tech-debt that is impacting system stability. This is something I usually advocate for, I think I've managed to get verbal approval at maybe half of my jobs and an actual commitment to dropping sprint work at like half of the half. 4) It has varied. The best was literally never, the worst was every single loving time I was on rotation at 3am because all of our customers were in a timezone where that was peak. 5) No, but I did manage to get the job where it always happened at 3am to agree that any time I spent on call counted double for time off the following day. So if I spent 3 hours resolving poo poo from 3-6am I would pop into Slack at 3PM the next day lol.
|
|
#
?
Nov 1, 2023 04:42
|
|
|
I have a lot of strong oncall opinions so here we go. My experience is from working at a company major tech companies, so YMMV if you're working at a tiny company with like two customers or a startup or something, but I suspect a lot of this is still applicable. Assuming you run a production service, your oncall is generally critical to keeping your poo poo running, and more importantly to you the individual, to keeping you from spending too much of your time doing incident followup / postmortem / whatever your company calls it. An ounce of prevention is worth a pound of cure and all that. Don't let yourself go 'eh, we'll figure it out.' Sit down, hammer out a process, write it down, and worst case you only use it a few times a year in emergencies, but you'll be glad you have it. Oncall is work, and your team should recognize it as such. IMO, oncall should get compensation time off if you're working outside of normal business hours. My current team comps half of your weekend oncall, which I think is a reasonable compromise because it's pretty quiet. During business hours, if your oncall is busy enough to impact your dev time, it needs to be prioritized over your development time and it needs to also be a clear separation; if you're oncall then that's your top priority , and your project/dev work is secondary and you shouldn't take any impact to your productivity from your oncall time, because you shouldn't be expected to be outputting dev work while oncall. (again, this assumes that your oncall takes up at least 50% of your time on average when oncall. Tiny shops/etc YMMV). The flip side of this is that oncall is work, and so the people participating in it should take it seriously; not because of some loyalty to your company, but because not taking it seriously fucks over your coworkers, who are human beings you should probably have a level of respect for. You should have a clear expectation on what your response time targets/etc are - not just to make some manager happy, but because being clear up front reduces the chance of drama and interpersonal bullshit. I had a guy who would just...randomly change the oncall at 2am because he was 'having problems with his internet' without a heads up or anything to whoever he cut it over to and later the same guy pulled the 'Oh well I was oncall but I was out at dinner so I couldn't take the call'. Some examples of those expectations: "Primary is available within 15 minutes to run an incident, start to finish" <-- This sets the expectation that primary needs to be near a working internet connection and have their laptop on them; if they want to go out to dinner but can run an incident from a hallway or their car? That's on them as long as they can do it. "Secondary needs to be available within 60 minutes of notification" <-- This makes it clear that secondary can't go camping/etc. You'd be surprised how often I've had to have that exact discussion with people on my team that no, you have to get coverage if you're literally going off the grid. Get coverage. "If secondary is not available, page the manager to find coverage" <-- another important thing, the guy needing help should not be on the hook for finding it if they're tied up in a thing already. Your management needs to be part of your escalation chain. (or an alternative escalation chain as appropriate to your team) Speaking of coverage, nothing looks worse than 'oh well we didn't get coverage for John going camping this weekend because he was secondary and we don't get called as secondary so when the primary lost power 10 minutes before the west coast datacenter caught fire, we had to spend hours trying to find someone." The point of having clear expectations and coverage / etc is to make it clear when you are and aren't available. If I'm not oncall and poo poo happens, then sure, if it's an absolute emergency I might be available, but I might also be six beers in and have my phone in a lake, and a good team isn't going to hold that against me. You should feel like you're getting fairly compensated for your oncall time; either as part of your salary because the oncall was a hiring expectation, or through other means (bonus pay, comp time, etc); this is a personal decision and something between you and your team, but you should not accept 'I'm making below market wage and I'm 24x7 oncall by myself' - walk from the fuckin' job and let them find someone willing to soak it. Sleep interruption ruins your life. Do not underestimate the impact of constant sleep interruption from a crappy oncall rotation. If the things waking you up are your problem, fix them. If you can't, make it your boss's problem. Constant sleep interruption makes your entire life worse and is not worth the cost. You're never going to avoid emergencies at any company that has a website, but being woken up once a year because some random poo poo broke is a far cry from my last team where we got paged 2.5x a night on average, and each engagement was an average length of 2-3 hours. That poo poo sucked, I wouldn't do it again for startup equity. There is no right answer for scheduling things like oncall. Find the solution that pisses everyone off the least, and you're probably good. Generally, people like predictable schedules because it helps them plan their life.
|
|
#
?
Nov 1, 2023 05:14
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. I got suckered into on-call between various departures, my being promoted and me knowing the most about the thing. > How often and for how long are you on-call? 12 hours a day, 7 days a month >What kind of duties are you expected to do for on-call? I mainly attend to pages that exclusively come from performance monitors/various alerts. I've worked on turning it into a relatively easy rotation of just passing the buck on slack to the right team. Occasionally, I'll need to really get into the weeds to figure out who hosed up if it's not easy to tell, but that's maybe a once a quarter occurrence. People message me sometimes for help because they see my name on rotation, but I either entirely ignore them or just link them to the channel they should go to during the next business day. We're pretty solid on the rotation being only for when poo poo hits the fan and the thing is down. > How often does something happen while you're on-call?(on or off hours) There are some false alerts I ack several times per rotation. I and the other primaries are very keen on eliminating false positives and tuning the stream of poo poo down to be mostly signal. I actually need to tag some team to resolve something probably every other rotation, but I don't do anything beyond that. A few times a year, I'm doing the dreaded sweating bullets live incident response. It's usually pretty bursty, we'll go through months of calm and then something that's been piling up eventually explodes and causes a noisy rotation. It's been a while though. The rotation used to be a lot more involved years ago, but we've been working on isolating concerns and getting teams to own their poo poo. > Do you get paid anything additional for on-call? No biceps crimes fucked around with this message at 05:42 on Nov 1, 2023 |
|
#
?
Nov 1, 2023 05:36
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. wilderthanmild posted:Is on-call mandatory for you? wilderthanmild posted:How often and for how long are you on-call? wilderthanmild posted:What kind of duties are you expected to do for on-call? wilderthanmild posted:How often does something happen while you're on-call?(on or off hours) wilderthanmild posted:Do you get paid anything additional for on-call?
|
|
#
?
Nov 1, 2023 17:06
|
|
|
Teams that have Stockholm syndrome for alerts are the worst. Alerts are actionable and something you need to be woken up for. Don't have stupid poo poo alert you. Don't be ok with being woken up at 3am for that flappy alert that just sometimes acts up.
|
|
#
?
Nov 2, 2023 01:33
|
|
|
luchadornado posted:Teams that have Stockholm syndrome for alerts are the worst. Alerts are actionable and something you need to be woken up for. Don't have stupid poo poo alert you. Don't be ok with being woken up at 3am for that flappy alert that just sometimes acts up. I've almost escalated to shouting before because I worked with teams that would just soak stuff like this and wouldn't fix it. I wasn't even the one waking up and I was furious.
|
|
#
?
Nov 2, 2023 01:51
|
|
|
I thought the whole point of SWE carrying the pager instead of ops was to motivate fixing stuff.
|
|
#
?
Nov 2, 2023 03:18
|
|
|
Paolomania posted:I thought the whole point of SWE carrying the pager instead of ops was to motivate fixing stuff. requires prioritization, time, and resources from management
|
|
#
?
Nov 2, 2023 03:35
|
|
|
Requires devs that don't go "hurr durr, if only there was a better way - this is a badge of honor!"
|
|
#
?
Nov 2, 2023 03:48
|
|
|
Paolomania posted:I thought the whole point of SWE carrying the pager instead of ops was to motivate fixing stuff. This might work in Very Large companies Smaller companies are worried about talent walking out the door and just shove it on to ops
|
|
#
?
Nov 2, 2023 04:23
|
|
|
Hadlock posted:This might work in Very Large companies Lol big companies just say "that alarm isn't in your deadlines list why are you trying to fix it?" Edit: To expand on this, I used to work at a particular BigTech company, and in a team that handled incident management, and had since Ye Olde Days when you had tier 1 getting all alarms, escalating to tier 2, and finally to the devs. Well, one day we made a change - there was no more escalation model, alarms went to the incident management team and the service team in parallel. For the majority of teams, this produced a very real effect to their false alarm rate, which was what we expected. But oh boy, a few of them. A few of them somehow would just...keep firing these false rear end alarms. They weren't a follow the sun team, so this wasn't 'ah some guy in Zurich's afternoon tea is ruined', this was like 'wow this is the uh...tenth time this month we've been on a call with y'all at 3am huh?'. I set aside time unprompted and did some decent analysis that at least one set of their alarms was complete garbage and that they could not only drop it entirely, but that they already had a second set of alarms that would have caught any real issues, and sent it over to them, and...nothing. I even escalated to their manager, who got pissy with me for interfering with his team. Surprise surprise that most of the quality engineers left that team over the next year or so. It was loving baffling to me and I can only imagine it was an incredibly toxic work environment. Falcon2001 fucked around with this message at 06:05 on Nov 2, 2023 |
|
#
?
Nov 2, 2023 05:00
|
|
|
Paolomania posted:I thought the whole point of SWE carrying the pager instead of ops was to motivate fixing stuff. In practice, the SWEs weren't knowledgable enough on SRE matters to fix that stuff, and they didn't want to invest learning, so whenever an emergency happened they'd just loop in an SRE who could mitigate the problem quickly. I saw the primary value of DevOps was that Dev got better empathy for Ops's job, and accordingly designed their software to fail less (better tests and higher quality code, don't YOLO deployments), and fail better (more useful logs, fallbacks, avoid cascading failures, etc). It did not result in less on-call for the SREs because all Devs got added to the on-call rota.
|
|
#
?
Nov 2, 2023 06:15
|
|
|
minato posted:I saw the primary value of DevOps was that Dev got better empathy for Ops's job, and accordingly designed their software to fail less Primary value of devops is to be the whipping boy and shame developers into writing better code 
|
|
#
?
Nov 2, 2023 06:49
|
|
|
Hadlock posted:Primary value of devops is to be the whipping boy and shame developers into writing better code The point was to cut costs by getting rid of the ops team and force the devs to do that job as well.
|
|
#
?
Nov 2, 2023 11:22
|
|
|
leper khan posted:The point was to cut costs by getting rid of the ops team and force the devs to do that job as well. Oh wow did you work at my last 1... 2.. 3... 4 jobs as well too? "This poo poo is mostly stable now, what are we still paying these guys for anyways??"
|
|
#
?
Nov 2, 2023 18:58
|
|
|
wilderthanmild posted:Curious of the broader experience with on call because I loving hate on call/pager duty/etc. 1. Yes, in every team I've been on we've all carried the pager. On my current team we have an escalation system where ICs and team managers are tier 1, senior managers are tier 2, and the product director is tier 3. Past teams have had a similar setup. 2. 1 week primary, 1 week secondary. Our rotation is comprised of all the teams in our product area so on-call is usually the 2 weeks every 5-6 months or so. I just checked PagerDuty and my last primary shift was August 2nd - August 29th and my next one is scheduled for February 14th to February 21st. 3. Handle pages, triage alerts, etc. We used to do a system where secondary would handle internal escalations but it just ended up being a "bug the secondary for tech support rather than reading the docs" so we moved that escalation process through product management and low and behold the number of requests dried up overnight. 4. Lately it's been worse for on-call folks because of issues mentioned in this thread. A lot of acking pages and waiting for error rates to fall without actually solving for the issue. We've got a persistent page around DB load during a particular release type where the resolution would be to tune a worker pool but lol no one seems to care. We also have a DB that's reaching txid wraparound that requires some auto-vacuum tuning but no one cares to fix it. It's maddening to me but not my horse, not my rodeo until I'm on call in which case I'll fix it. 5. Lol I wish.
|
|
#
?
Nov 3, 2023 14:26
|
|
|
Has anyone come across anything out there about optimal team size and experience range? I'm having a hard time wrapping my head around my current team's struggles - it'd be great to have a starting point and some general thoughts on these topics from trustworthy sources.Blinkz0rz posted:We also have a DB that's reaching txid wraparound that requires some auto-vacuum tuning but no one cares to fix it. 
|
|
#
?
Nov 3, 2023 23:38
|
|
|
luchadornado posted:Has anyone come across anything out there about optimal team size and experience range? I'm having a hard time wrapping my head around my current team's struggles - it'd be great to have a starting point and some general thoughts on these topics from trustworthy sources.
|
|
#
?
Nov 4, 2023 05:00
|
|
|
wilderthanmild posted:Yeah, but it's more fun being the guy who created all the tech debt and then left. You get to have fun and just code, future devs get to blame you for all problems. It's win win. Dragging up an old topic, this is kind of where I'm landing. This quote really resonated with me
|
|
#
?
Nov 4, 2023 08:50
|
|
|
You, too, can build a FastCGI in awk and elm and then gently caress off into the severance sunset. Again and again.
|
|
#
?
Nov 4, 2023 13:20
|
|
|
This is a very good post. I don't care for the availability rules, so I avoid on-call positions. Until Microsoft recruiter just flat out lied to me during call about whether a position is on-call or not  I ended up leaving in the probationary period after letting them pay me for further job search, because gently caress that.
|
|
#
?
Nov 4, 2023 15:52
|
|
|
Startyde posted:You, too, can build a FastCGI in awk and elm and then gently caress off into the severance sunset. Again and again. people keep saying things like "it's all about the business needs" and this is a logical conclusion
|
|
#
?
Nov 4, 2023 20:34
|
|
|
I'm considering starting a side gig as a tutor. Has anyone here done this before? Any advice or suggestions for ways to prepare? I have plenty of professional experience, but my teaching experience is pretty much "was an L5, wrote a lot of documentation, taught people to play board games".
|
|
#
?
Nov 5, 2023 03:27
|
|
|
TooMuchAbstraction posted:I'm considering starting a side gig as a tutor. Has anyone here done this before? Any advice or suggestions for ways to prepare? I have plenty of professional experience, but my teaching experience is pretty much "was an L5, wrote a lot of documentation, taught people to play board games". It can be super rewarding but it might be harder than you think. It's not going to be "Here are the cool parts of search algorithms" it will be "Here is the process you need to go through, let me explain it 5 times until it sinks in". It isn't total drudgery, but consistency and simplicity tend to be more helpful than fun facts. But my experience is from helping high school students decades ago.
|
|
#
?
Nov 5, 2023 03:57
|
|
|
TooMuchAbstraction posted:I'm considering starting a side gig as a tutor. Has anyone here done this before? Any advice or suggestions for ways to prepare? I have plenty of professional experience, but my teaching experience is pretty much "was an L5, wrote a lot of documentation, taught people to play board games". Like as part of a formal school system or? Most k-12 school systems will let you sign up to substitute teach with any college degree. Depending on where you are, they might be willing to hire you as a full time teacher as well, even without any formal training in teaching. I can't really recommend either. Teaching a full class of 30+ kids is a lot more babysitting and less conveying knowledge than the sort of 1-1 mentoring more common in industry. StumblyWumbly posted:It can be super rewarding but it might be harder than you think. It's not going to be "Here are the cool parts of search algorithms" it will be "Here is the process you need to go through, let me explain it 5 times until it sinks in". It isn't total drudgery, but consistency and simplicity tend to be more helpful than fun facts. Teaching people to play board games is pretty relevant really. You explain the rules (or the process or technique), but it doesn't really sink in until after playing the game (or trying to do something themselves). It usually takes several games/attempts for people to fully absorb anything. And even then edge cases are likely to trip people up.
|
|
#
?
Nov 5, 2023 04:54
|
|
|
StumblyWumbly posted:It can be super rewarding but it might be harder than you think. It's not going to be "Here are the cool parts of search algorithms" it will be "Here is the process you need to go through, let me explain it 5 times until it sinks in". It isn't total drudgery, but consistency and simplicity tend to be more helpful than fun facts. Yeah, that tracks, and I'm pretty sure I'm OK with that. LLSix posted:Like as part of a formal school system or? Yeah, I'm confident I don't want to do classroom work. That does not sound nearly as fun, plus it's a much bigger time commitment. I'm imagining something that I could devote roughly one day's worth of work hours to per week. quote:Teaching people to play board games is pretty relevant really. You explain the rules (or the process or technique), but it doesn't really sink in until after playing the game (or trying to do something themselves). It usually takes several games/attempts for people to fully absorb anything. And even then edge cases are likely to trip people up. I thought that might be relevant! And yeah, needing to "play through" a few times before stuff really sinks in makes sense. Thanks for the insights! I guess really what's giving me anxiety is that I'll have to have a first client at some point, where I have no actual direct experience. I don't want to gently caress it up.
|
|
#
?
Nov 5, 2023 21:39
|
|
|
TooMuchAbstraction posted:Yeah, that tracks, and I'm pretty sure I'm OK with that. If you've mentored / helped people at work before you're not going in totally blind, in my opinion ( I haven't worked as a tutor, so take this all with a grain of salt). As long as the person being tutored is even remotely interested in the subject, the worst you'll probably run into is spending more time than expected on tutoring and prep. I would say that prep is going to be a big part of it too, especially as you get started. It's probably on a case-by-case basis but I would expect to need to build a lesson plan and also have a general plan for going through it, which will be rough for the first client or so. I would just not double book yourself or anything.
|
|
#
?
Nov 6, 2023 18:20
|
|
|
So I could ask the laptop thread about this but this might be a better audience My boss is giving me some budget for home office stuff, I guess I'm gonna be the team guinea pig on the M3 MBP and thunderbolt 4 connectors Does anyone have an M3 MBP and what docking station are you using with it? In particular what's your solution with dual external displays? I have a thinkpad branded thunderbolt 3 docking station but it's always kind of struggled with doing apple hardware stuff consistently and without being flaky Thinking about going dual Dell Ultrasharp 24" displays
|
|
#
?
Nov 6, 2023 23:28
|
|
|
Hadlock posted:Thinking about going dual Dell Ultrasharp 24" displays I have an Ultrasharp Curved 34" and it's the best monitor and best docking station I've ever owned. It's also like $700 though so anything short of that would have been a massive disappointment. I am using it with an M1, but not sure that makes any difference here.
|
|
#
?
Nov 6, 2023 23:52
|
|
|
Hadlock posted:So I could ask the laptop thread about this but this might be a better audience I didn't think the M3 was actually available yet? Anyway, I have an M1 Max MBP, and I've just got a new Caldigit TS4 that works with it, and two Dell Ultrasharp 25" displays. It replaced an older Calgiti Pro dock that pooped out recently. While I was waiting for that, I temporarily used an old Lenovo thinkpad TB 3 docking station, but it did not work with both monitors. The reason, after I dug around for a while, is that even though the Lenovo docking station had two DisplayPort ports, internally it used DisplayPort Multi-Stream Transport (what DP uses for daisy chaining) to support both on a single connection, and Apple does not support that. So maybe you have a newer Lenovo station, but I'd definitely figure that out because I think there's good odds it won't work with a Macbook. The TS4, while pricey, has been working great.
|
|
#
?
Nov 7, 2023 03:20
|
|
|
Last two stops have given me an M1 and then an M2 pro system. I have an Anker TB4 dock that ran ~$250 that has been fantastic, but the pro chips support dual externals natively. Most recent stop let me buy 2 business class monitors to replace my old 24" Samsungs from 2011. I got two LG Dual Up monitors on sale for $500 each, and they're so good.
|
|
#
?
Nov 7, 2023 16:20
|
|
|
|
| # ? May 25, 2024 23:41 |
|
|
Hadlock posted:Freeze over I dunno about anyone else but since the first day of Q4 seems like people can't hire fast enough Guy I talked to today they're hiring six additional people on that one team alone they just don't have the time budget to onboard that many people so quickly, and I believe them
|
|
#
?
Nov 10, 2023 05:56
|
|