The BBC frames the shelter directive itself as the newsworthy event, emphasizing that NASA and Roscosmos closed the hatch between segments and restricted crew movement. This posture is distinct from the routine sealant patching cosmonauts have done for years, indicating the leak has crossed a threshold that warranted active crew protection.
By submitting the BBC live-coverage story to HN with the framing 'astronauts told to shelter,' the submitter highlights the shelter order as the key signal. The 401-point score and 251 comments suggest the community read this as a meaningful escalation rather than business as usual.
The editorial argues the PrK cracks aren't a discrete failure to be patched but the visible symptom of a 25-year-old pressure vessel cycling through thermal and mechanical loads beyond its qualification. The leak rate has stepped up at least four times since 2019 despite continuous sealant application, supporting the view that mitigation can only slow, not stop, the underlying fatigue process.
The editorial emphasizes that NASA attributes the cracks to metal fatigue and possible micrometeoroid contribution, while Roscosmos has publicly blamed docking vibrations. This isn't an academic dispute — it determines whether the leak is sealable or merely manageable, and the NASA OIG flagged it as a top-tier safety risk in September 2024 precisely because the partners can't agree on what's actually happening.
The editorial points to cargo manifests for upcoming Progress and Cygnus flights, which include extra sealant and pressure-monitoring instrumentation, as a more honest indicator of partner concern than the calm public messaging. How agencies budget logistics internally reveals their actual threat assessment in a way press conferences don't.
NASA and Roscosmos directed the seven-person ISS crew to shelter in the Russian Orbital Segment while ground teams worked the latest escalation of a long-running air leak in the Zvezda service module's PrK transfer tunnel. The hatch between the U.S. segment and the affected Russian compartment was closed during repair operations, with crew movement restricted until pressure telemetry stabilized. No injuries. No emergency evacuation. But the shelter order itself is the news — that's not a routine maintenance posture.
The leak in PrK has been actively tracked since September 2019, and the loss rate has stepped up at least four times since then, most recently in early 2024 when it briefly hit roughly 3.7 pounds of air per day before partial mitigation. Russian cosmonauts have spent the intervening years applying sealant to a growing constellation of hairline cracks on the interior walls of the tunnel. NASA's Office of Inspector General flagged the leak as a top-tier safety risk in a September 2024 report. The agencies disagree on the root cause — NASA suspects metal fatigue compounded by internal/external stress cycling and possible micrometeoroid contribution; Roscosmos has at times publicly attributed it to vibration from docking events. The disagreement matters because it dictates whether the leak can be sealed or only slowed.
The station currently hosts NASA, Roscosmos, JAXA, and commercial crew under Expedition 73. Cargo manifests for the next two Progress and Cygnus flights include additional sealant and pressure-monitoring instrumentation, which tells you how the partners are budgeting for this internally even when the public statements stay calm.
The ISS is the most expensive object humans have ever built and it is, quietly, aging out. The leak isn't a discrete defect — it's the visible symptom of a 25-year-old pressure vessel cycling through thermal and mechanical loads it was never qualified to survive this long. Zvezda launched in 2000. Its original design life was 15 years. We are now in year 26, with retirement officially set for 2030, and SpaceX has the $843M contract for the U.S. Deorbit Vehicle that will eventually drag the whole stack into the Pacific.
The interesting question isn't whether the current leak gets patched. It will. The question is what the leak tells you about the margin between "operational" and "unsafe" on a structure that has no replacement waiting in the wings. Axiom, Vast, Voyager, and Blue Origin are all building commercial successor stations, but none of them will have crewed continuity with ISS. Axiom's first module slips to no earlier than 2027. Vast's Haven-1 is targeting a 30-day single-crew demo, not a permanent platform. If ISS goes from "degraded but flyable" to "unsafe to occupy" before 2028, the U.S. loses continuous human presence in LEO for the first time since November 2000 — a 28-year streak.
The community reaction on the HN thread split predictably between "this is a normal part of operating aging hardware" and "this is the canary." Both are right. NASA has flown the Shuttle with known foam-shedding issues, ISS with known ammonia leaks, and Soyuz with known coolant loop failures. Operating degraded systems with bounded risk is what spaceflight is. But the bound on Zvezda is moving. The OIG report explicitly noted that NASA and Roscosmos have not aligned on the leak-rate threshold that would trigger evacuation of the Russian segment — a procedural gap that should have been closed two years ago.
There's also a geopolitical layer worth naming. The Russian segment provides ISS propulsion and reboost. The U.S. segment provides power and most of the science. Neither side can fly the station alone. Roscosmos has publicly threatened ISS withdrawal multiple times since 2022 without follow-through, but a Zvezda failure mode that forces permanent isolation of the ROS would functionally end the partnership early — and force NASA into an awkward conversation about whether Cygnus reboost (demonstrated in 2022 and 2025) can carry the station to 2030 without Russian propulsion.
For working engineers this is a long-cycle infrastructure story, not a deploy-tomorrow story, but the lessons port. If you operate any system past its design life, you are inheriting a fatigue budget you didn't model and can't fully observe. The ISS leak telemetry is the best-instrumented version of this problem in human history — pressure decay tracked to the gram, thermal cycles logged for decades — and the partners still can't agree on root cause. Your production system has worse instrumentation and the same problem.
If you're in the commercial space supply chain — propulsion, life support, attitude control, sealants, sensors — the next 36 months are when the deorbit vehicle and the successor stations procure. SpaceX's USDV contract is one program; the CLD (Commercial LEO Destinations) awards to Axiom, Voyager/Starlab, and Blue Origin's Orbital Reef are the others. Lead times on flight-qualified hardware are 18-24 months. The bidder lists are largely set but subcontract slots aren't.
If you work in critical-infra observability more broadly, the Zvezda situation is a clean case study for the limits of "trend the metric." The leak rate has been monotonically increasing in step functions for six years. Everyone sees it. The disagreement is on what causes the steps and what the next step will look like. Monitoring without a mechanistic model gets you arguments, not decisions.
The shelter order will lift, the sealant will hold for some weeks or months, and the leak rate will step up again. The real watch item is whether NASA and Roscosmos publish an aligned evacuation threshold before the next step — and whether Cygnus reboost gets formally added to the station's nominal propulsion plan. If both happen quietly in the next two quarters, ISS makes it to 2030. If neither does, the over/under on the deorbit timeline moves into 2028.
> After multiple inspections and sealant applications, Nasa reported in January that pressure readings suggested a stable configuration had been reached - though there remained uncertainty about whether the leak had truly been sealed or whether air was simply escaping elsewhere.I'm clearly n
Maybe someone who knows more about the ISS than I do can answer this:Naively, I would assume that there are airlocks between the different sections of the ISS. I would also assume that they would close these airlocks while doing the kind of work they are doing to repair the leaks.So, assuming I'
Can't they just get things out of the module and paint it fresh? Maybe with some special paint, or with several layers of a paint?Obviously they can't, it looks like an obvious solution they couldn't have missed. But I wonder why it is impossible to do.
They should keep some FlexSeal up there !
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
I found this interesting: NASA RELL (Robotic External Leak Detector) [1]. "NASA’s Robotic External Leak Locator (RELL) is a robotic, remote-controlled tool that helps mission operators detect the location of an external leak and rapidly confirm a successful repair. … Two instruments working in