How to Monitor Live Streams for Optimal Ad Server Performance
From an observability standpoint and eyes on glass, what are the key metrics to watch to determine if ad insertion is going smoothly during a live stream? Altitude TV SVP of operations and engineering Dave Zur, Sargeway LLC owner Sarge Sargent, Qualabs chief solutions architect David Hassoun, and FanServ VP of publisher operations C.J. Leonard discuss challenges and best practices for monitoring live streaming ad performance in this clip from Streaming Media Connect 2026.
Garbage In, Garbage Out
Leonard asks, “From an operational standpoint on the engineering side, what monitoring or observability signals should teams prioritize?” If there’s a problem that someone hasn’t flagged from the ad server side, she wonders, what can be done?
Zur notes that the encode is the first step—checking video quality and integrity—then, asking if the manifest has been degraded in any way. “I mean, we’re only as good as our CDNs and our players at the end of the day,” he says. “We don’t have the traditional broadcast infrastructure of eyes on glass. In many of these cases, somebody might be halfway around the world in a different time zone. It’s yesterday or it’s tomorrow, whatever it might be, and we’re interacting with them live regarding problems that are popping up. That’s been maybe one of the bigger challenges as far as the monitoring.”
If there’s a resolution problem, “you can only do so much,” Zur continues. He subscribes to the garbage in, garbage out rule—or quality in, quality out—but, he says with a laugh, there are days with “quality in” that “we get a little garbage on the backside.”
Sargent agrees: “Garbage in, garbage out is definitely true to the system.” Given his experience in both video engineering and ad tech, he suggests modeling what Disney does: “[T]hey actually do monitor encode because to your point, if it’s [garbage] coming in, we’re definitely [getting garbage] coming out. We’re also monitoring the SCTE[-35] signals. So we’re making sure the ad breaks are proper in place. Not only are they on time, but are they of the right type? And for Peacock’s workflows, we actually did embed custom code into the SCTE signal. So we would send messaging that described the ad break so the ad server had more detail so it can do different things.”
Multiple People, Multiple Devices
Sargent uses Slack or Teams for real-time communications with operation systems and support systems, which is working for him. “I believe, talking to some of our peers, [that] a lot of them still use Slack to do real-time communications, but you’ve got to have the eyes on glass. That’s kind of mandatory for a live event, especially tentpole events.”
Multiple people need to be on glass, he says, from an in-house data center or a NOC (network operations center) or DNOC (deployable network operations center). “[F]or example, my setup, I have multiple TVs and multiple devices running. And same for my teams. We want to have multiple devices. So you’re looking at the ad break on a Roku versus a tvOS and definitely multiple desktop versions, different browsers. So you’re getting a full range of what’s happening in the stream,” Sargent explains.
Monitoring Delivery Time
Hassoun says that when monitoring large events, he needs to see the metrics about the time from ad request to received. “If that starts to grow, and how’s that handling with growing audience, that’s really, really important,” he notes. “That’s an indicator if you’re having ad decisioning services issues, as well as a little bit on the delivery, but the delivery itself is very minor. They’re small files, but high value.” He uses Conviva to see how the actual ads are being delivered, he says. “If I was just monitoring a Conviva dashboard for live, I could tell you exactly when an ad break was without ever looking at a screen just by looking at the results because the ads always get delivered slower. They’re not as well cached.”
Hassoun adds that he looks at delivery and how it’s affecting buffering. “And once again, if you have lower latency, you have less amount of buffer, does that create drift now to content? And those are other factors within there,” he asserts. “And then to take a little bit further on what [Zur] and [Sargent] were talking about, it’s a matter of not just eyes on glass, which is super important, and also the encoding, but looking at what happens with those ads.”
Getting Pieces of the Big Picture
Speaking to bad experiences he’s had with ads hurting the viewing experience, Hassoun describes machine lags, improper encoding, and other problems. “So having those eyes on glass and being able to review those ads—especially if you can’t review them all ahead of time, if they are going to be programmatic—but [having] many screens, many devices as [Sargent] said,” can make a big difference, Hassoun notes. It helps isolate issues so problems can be solved.
Hassoun sums up, “So it’s a matter of looking at all your network delivery, looking at your performance metrics from your player, your QOE [quality of experience], QOS [quality of service], and trying to also have that data around your ad services gives you a little bit of that picture.” It doesn’t provide the entire picture, he says, because “this is really hard to monitor and see unless you have, as [Zur] mentioned, a node in every location and region and variance and personalization, [etc.], it becomes very challenging to have high confidence in everything.”
Join us May 12–14, 2026 for more thought leadership, actionable insights, and lively debate at Streaming Media Connect 2026! Registration is open!
Related Articles
With sports that aren't orchestrated for built-in TV ad breaks and a traditional broadcast run of show, SGAI provides critical advantages for monetizing sports content without pulling fans out of the action and causing them to miss key moments in the game. FloSports Product Manager, Ad Tech Adair Lyden explains these and other reasons why FloSports is embracing server-guided ad insertion (SGAI) as a good fit for their content and programming model and provides an early look at how it's working for them in this conversation with RingDigital's Brian Ring at Streaming Media Connect 2026.
07 Apr 2026
Streaming Media Connect February 2026 was all about the business and tech of ad-supported streaming and featured exclusive keynote fireside chats with Michelle Auguste of the NBA, Michelle Munson of Eluvio, and Kimmy Li of GAMELOOP and a slate of expert panels panels packed with speakers from YouTube, Disney, DAZN, Fubo, Fremantle, FloSports, Fuse, LG, Philo, Bitmovin, CacheFly, Tavant, and more. Check out a playlist with Streaming Media Connect sessions on Streaming Media's YouTube channel to catch the sessions you missed and revel in the ones you want to relive through the magic of VOD.
26 Feb 2026
VAST (video ad serving template) is an XML-based standard developed by the IAB Tech Lab that defines how video players and ad servers communicate to deliver video ads. It is designed to allow for the consistent and dynamic delivery of video ads across different platforms and devices. But in the ever-shifting and fast-evolving world of adtech, has VAST grown long in the tooth? Ring Digital's Brian Ring, SVTA's David Hassoun, Eventually a Castle's Andrew Baritz, and IAB Tech Lab‘s Raymond Holton discuss the issue in this clip from May's Streaming Media Connect.
22 Jul 2025
Reality Software's Nadine Krefetz, Sinclair's Jason Justman, and fuboTV's Geir Magnusson discuss the importance of SCTE 35 to server-side ad insertion in this clip from their panel at OTT Leadership Summit at Streaming Media East 2019.
20 Sep 2019