[HN Gopher] OpenTelemetry Is Great, but Who the Hell Is Going to...
       ___________________________________________________________________
        
       OpenTelemetry Is Great, but Who the Hell Is Going to Pay for It?
        
       Author : thunderbong
       Score  : 36 points
       Date   : 2025-06-30 18:28 UTC (4 hours ago)
        
 (HTM) web link (www.adatosystems.com)
 (TXT) w3m dump (www.adatosystems.com)
        
       | denysvitali wrote:
       | I don't think the comparison is correct. For sure OTEL adds some
       | overhead, but if you're ingesting raw JSON data, then even with
       | the overhead it's probably going to be reduced since internally
       | the system talks OTLP - which is often (always?) encoded with
       | protobuf and most of the time sent over via gRPC.
       | 
       | It's then obviously your receiver end's job to take the incoming
       | data and store it efficiently - grouping it by resource
       | attributes for example (since you probably don't want to store 10
       | times the same metadata). But especially thanks to the
       | flexibility of adding all the surrounding metadata (rather than
       | just shipping the single log line), you can do magic thinks like
       | routing metrics to different tenants / storage classes or drop
       | them.
       | 
       | Having said that, OTEL is both a joy and an immense pain to work
       | with - but I still love the project (and still hate the fact that
       | every release has breaking changes and 4 different version
       | identifiers).
       | 
       | Btw, one of the biggest win in the otel-collector would be to use
       | the new Protobuf Opaque API as it will most likely save lots of
       | CPU cycles (see https://github.com/open-telemetry/opentelemetry-
       | collector/is...) - PRs are always welcome I guess.
        
       | maplemuse wrote:
       | The part about SNMP made me laugh. I remember integrating SNMP
       | support into an early network security monitoring tool about 25
       | years ago, and how it seemed clunky at the time. But it's
       | continued to work well, and be supported all these years. It was
       | a standard, but with very broad tool support, so you weren't
       | locked into a particular vendor.
        
         | blinded wrote:
         | smmp-exporter ftw
        
           | rbanffy wrote:
           | And, for a lot of things, it's quite sufficient.
           | 
           | I used Munin a lot as well in the 2005-2010 timeframe. Still
           | do as a backup (for when Prometheus, Grafana, and Influxdb
           | conspire against me) on my home lab.
           | 
           | Usually the 15 minute collection interval is just fine. One
           | time though I had an issue with servers that were just fine
           | and, then, crashed and rebooted with no useful metrics
           | collected between the last "I'm fine" and the first "I'm fine
           | again".
           | 
           | At that point we started collecting metrics (for only those
           | servers) every 5 seconds, and we figured out someone
           | introduced a nasty bug that took a couple weeks of uptime to
           | run out of its own memory and crash everything. It was a fun
           | couple days.
        
       | dboreham wrote:
       | Uhhh. The point of OTel is that you can host it yourself. And
       | should do imho unless you're part of a VC money laundering scheme
       | where they want to puff up NR or DD or whoever portfolio company
       | numbers.
        
         | rbanffy wrote:
         | You should always think about how much it'll cost for you to
         | roll out and maintain something vs how much it would cost to
         | buy the service from a vendor.
         | 
         | Chances are your volumes are low enough it will be actually
         | cheaper to run with something like New Relic or Datadog. When
         | the monthly bill starts reaching 10% of what a dedicated person
         | would cost, it's time to plan your move to self-hosted.
        
           | mdaniel wrote:
           | > it's time to plan your move to self-hosted.
           | 
           | No, it's always time to plan the move to self hosted, and
           | just occasionally choose someone else to be the "self."
           | Because once a proprietary vendor gets in the stack, evicting
           | them is going to be a _project_
           | 
           | I'm aware that this doesn't split cleanly down the "saas only
           | feature" or the evil "rug pull" axes, but I'd much rather say
           | "I legitimately tried to allow us to eject from the walled
           | garden and the world changed" versus "whaddya mean non-
           | Datadog?"
        
         | jsight wrote:
         | In my experience, the people willing to pay the most to not
         | host it themselves are often the big companies that are long
         | past VC money.
         | 
         | They'll gladly pay someone to do it and have a big team of
         | engineers and planners to support the outsourcing.
         | 
         | Efficiency isn't what bigco inc is about.
        
           | xyzzy123 wrote:
           | BigCos have seen teams come and go, whole departments
           | slaughtered by reorgs. They have seen weird policy changes,
           | political battles, personal beefs and bad managers that
           | trigger waves of attrition.
           | 
           | They know that even if you have the capacity to run something
           | internally today, that is a delicate state of affairs that
           | could easily change tomorrow.
        
       | hermanradtke wrote:
       | New Relic, Datadog, etc are selling their original offering but
       | now with otel marketing.
       | 
       | I encourage the author to read the honeycomb blog and try to grok
       | what makes otel different. If I had to sum it up in two points:
       | 
       | - wide rows with high cardinality
       | 
       | - sampling
        
       | stego-tech wrote:
       | Excellent critique of the state of observability, especially for
       | us IT folks. We're often the first - and last, until the bills
       | come - line of defense for observability in orgs lacking a
       | dedicated team. SNMP Traps get us 99% of the way there with
       | anything operating in a standard way, but OTel/Prometheus/New
       | Relic/etc all want to get "in the action" in a sense, and hoover
       | up as much data points as possible.
       | 
       | Which, sure, if you're willing to pay for it, I'm happy to let
       | you make your life miserable. But I'm still going to be the Marie
       | Kondo of IT and ask if that specific data point brings you joy.
       | Does having per-second interval data points actually improve
       | response times and diagnostics for your internal tooling, or does
       | it just make you feel big and important while checking off a box
       | somewhere?
       | 
       | Observability is a lot like imaging or patching: a necessary
       | process to be sure, but do you _really_ need a Cadillac Escalade
       | (New Relic /Datadog/etc) to go to the grocery store when a Honda
       | Accord (self-hosted Grafana + OTel) will do the same job more
       | efficiently for less money?
       | 
       | Honestly regret not picking the Observability's head at BigCo
       | when I had the chance. What little he showed me (self-hosted
       | Grafana for $90/mo in AWS ECS for the corporate infrastructure of
       | a Fortune 50? With OTel agents consuming 1/3 to 1/2 the resources
       | of New Relic agents? Man, I wish I had jumped down that specific
       | rabbit hole) was amazingly efficient and informative. Observation
       | done right.
        
         | rbanffy wrote:
         | > But I'm still going to be the Marie Kondo of IT and ask if
         | that specific data point brings you joy.
         | 
         | There seems to be a strong "instrument everything" culture
         | that, I think, misses the point. You want simple metrics
         | (machine and service) for everything, but if your service gets
         | an error every million requests or so, it might be overkill to
         | trace every request. And, for the errors, you usually get a
         | nice stack dump telling you where everything went wrong (and
         | giving you a good idea of what was wrong).
         | 
         | At that point - and only at that point, I'd say it's worth to
         | TEMPORARILY add increased logging and tracing. And yes, it's OK
         | to add those and redeploy TO PRODUCTION.
        
           | Nextgrid wrote:
           | > but do you really need a Cadillac Escalade (New
           | Relic/Datadog/etc) to go to the grocery store
           | 
           | Depends if your objective is to go to the grocery store or
           | merely showing off going to the grocery store.
           | 
           | During the ZIRP era there was a financial incentive for
           | everyone to over-engineer things to justify VC funding rounds
           | and appear "cool". Business profitability/cost-efficiency was
           | never a concern (a lot of those business were _never_ viable
           | and their only purpose was to grift VC money and enjoy the
           | "startup founder" lifestyle).
           | 
           | Now ZIRP is over, but the people who started their career
           | back then are still here and a lot of them still didn't get
           | the memo.
        
             | stego-tech wrote:
             | > During the ZIRP era there was a financial incentive for
             | everyone to over-engineer things to justify VC funding
             | rounds and appear "cool".
             | 
             | Yep, and what's worse is that...
             | 
             | > Now ZIRP is over, but the people who started their career
             | back then are still here and a lot of them still didn't get
             | the memo.
             | 
             | ...folks let go from BigTech are filtering into smaller
             | orgs, and the copy-pasters and "startup lyfers" are
             | bringing this attitude with them. I guess I got lucky
             | enough to start my interest in tech before the dotcom
             | crash, my career just before the 2008 crash, and finished
             | my BigTech tenure just after COVID (and before the likely
             | AI crash), and thus am always weighing the costs versus the
             | benefits and trying to be objective.
        
               | Nextgrid wrote:
               | > folks let go from BigTech are filtering into smaller
               | orgs, and the copy-pasters and "startup lyfers" are
               | bringing this attitude with them
               | 
               | Problem is, not all of them are even doing this
               | _intentionally_. A lot actually _started_ their career
               | during that clown show, so for them this is normal and
               | they don 't know any other way.
        
           | mping wrote:
           | On paper this looks smart, but when you hit a but that
           | triggers under very specific conditions (weird bugs happen
           | more often as you scale), you are gonna wish you had tracing
           | for that.
           | 
           | The ideal setup is that you trace as much for some given time
           | frame, if your stack supports compression and tiered storage
           | it becomes cheap er
        
           | prymitive wrote:
           | > There seems to be a strong "instrument everything" culture
           | 
           | Metrics are the easiest way to simply expose your application
           | internal state and then, as a maintainer of that service,
           | you're in nirvana. And even if you don't go that far you're
           | likely to be an engineer writing code and when it comes time
           | to add some metrics why wouldn't you add more rather than
           | less, and once you have all of them why not adding all
           | possible labels? And in the meantime your Prometheus server
           | is in a crash loop because it run if of RAM, but that's not a
           | problem visible to you. Unfortunately there's a big gap in
           | understanding between a code editor writing instrumentation
           | code and the effect in resource usage on the other end of
           | your observability pipeline.
        
             | sshine wrote:
             | I can only say, I tried to add massive amounts of data
             | points to a fleet of battery systems once; 750 cells per
             | system, 8 metrics per cell, one cell every 20 ms. It became
             | megabits per second, so we only enabled it when engaging
             | the batteries. But the data was worth it, because we could
             | do data modelling on live events in retrospect when we were
             | initially too busy fixing things. Observability is a super
             | power.
        
         | jsight wrote:
         | >Observability is a lot like imaging or patching: a necessary
         | process to be sure, but do you really need a Cadillac Escalade
         | (New Relic/Datadog/etc) to go to the grocery store when a Honda
         | Accord (self-hosted Grafana + OTel) will do the same job more
         | efficiently for less money?
         | 
         | The way that I've seen it play out is something like this:
         | 1. We should self host something like Grafana and otel.
         | 2. Oh no, the teams don't want to host individual instances of
         | that, we should centralize it!         (2b - optional, but
         | common, Random team gets saddled with this job)       3. Oh no,
         | the centralized team is struggling with scaling issues and the
         | service isn't very reliable. We should outsource it for 10x the
         | cost!
         | 
         | This will happen even if they have a really nice set of
         | deployment infrastructure and patterns that could have allowed
         | them to host observability at the team level. It turns out,
         | most teams really don't need the Escalade, they just need some
         | basic graphs and alerts.
         | 
         | Self hosting needs to be more common within organizations.
        
       | pojzon wrote:
       | Difference between OTel and other previous standards is that OTel
       | was created by "modern" engineers that dont care about resource
       | consumption or dont even understand it. Which is funny because
       | thats what the tool is about.
       | 
       | So yea, cost of storage and network traffic is only going to
       | balloon.
       | 
       | There is room for improvements and I can already see new projects
       | that will most likely gain traction in upcoming years.
        
         | growse wrote:
         | One of the biggest fallacies I see in this space is people
         | looking at an observability standard like otel and thinking "I
         | must enable all of that".
         | 
         | You really don't have to.
         | 
         | Throw away traces. Throw away logs. Sample those metrics. The
         | standard gives you _capabilities_ , it doesn't force you to use
         | them. Tune based on your risk appetite, constraints, and needs.
         | 
         | My other favourite retort to "look how expensive the
         | observability is" is "have you quantified how expensive _not_
         | having it is ". But I reserve that one for obtuse bean counters
         | :)
        
       | cortesoft wrote:
       | Setting up a self hosted prometheus and grafana stack is pretty
       | trivial when starting out. I run a Cortex cluster handling
       | metrics for 20,000 servers, and it requires very little
       | maintenance.
       | 
       | Self-hosting metrics at any scale is pretty cost effective.
        
       | ramon156 wrote:
       | All I want is spanned logs in JS. Why do I need OTEL? Why can't
       | pino do this for me?
        
       | ris wrote:
       | The logging examples given don't appear to be too different to
       | what any structured & annotated logging mechanism would give you.
       | On top of that it's normally encoded with grpc, so that's already
       | one-up on basic json-encoded structured logs.
       | 
       | The main difference I see with otel is the ability to repeatedly
       | aggregate/decimate/discard your data at whatever tier(s) you deem
       | necessary using opentelemetry-collector. The amount of data you
       | end up with is up to you.
        
       ___________________________________________________________________
       (page generated 2025-06-30 23:01 UTC)